Description of the ESL 175
|Type:||Full range Wire Stator ESL|
|Initial designer:||E. Fikier (full credits for the design go to E. Fikier but since I changed some of his specifications I think its more appropriate to say initial design)|
|ESL size:||1600 x 260 x 80 mm|
|HV power supply:||7500 V|
|Amplifier power:||min. 40 W|
|Sensitivity:||82 (88) dB (1/W/1 m) with an audio transformer 1:50 (1:150)|
|Freq. responce:||60 Hz...20 kHz (audio transformer 1:50); 40 Hz...20 kHz (audio transformer 1:150)|
For an exact description of the ESL 175 I have to refer to the book of E. Fikier.
I have made the same electrostatic loudspeaker (ESL) as Hans Zeeuwe. This ESL (ESL 175) has been designed by E.Fikier and the original design, together with a lot of other useful information can be found in the book 'Elektrostatische luidsprekers, zelf bouwen of zelf kopen' (Dutch), written by E. Fikier, published by Elektuur.
This ESL uses tensioned wires, which are insulated with PVC, as stators. Just like Hans, I had problems to build them properly. So the first one I build didn't work very well. I started to think about the design and contacted the designer. This resulted in the use of other materials and construction methods. I will describe this modifications and give some further advice. A precise description of how to build the ESL 175 is not given here. For this, I refer to the book mentioned above. Some information will also be useful when making other ESLs than the ESL 175.
Construction of the stators. In the ESL 175, the electrostatic element consists of two wooden (stator-) frames in which small PVC tubes (16 mm) are glued. The insulated and tensioned copper wires are glued on this tubes forming a stator. It can be difficult to mount all the tubes in exactly the same plane. But this is easy when the frame is laid down on a flat surface. The tubes are glued to the frame and by pressing the tubes down, glue at the bottom will be pushed away. Don't forget to lay a thin piece of paper/plastic under the tubes/frame, otherwise the tubes will be glued with the flat surface too. A problem that still exists is that all the tubes have to be cut precisely to fit in the wooden frame. Moreover, because the tubes are hollow, the contact surface (glue!) with the frame is limited and if they are not cut straightway, this may even be worse. You may solve this in some respect by 'closing' the tubes with some solid epoxy, but this is a lot of work and binding with the PVC can be bad. Therefore, the following method is useful: In the original design, one stator frame consists of two wooden frames (18 mm thick each) of equal size. Instead of using two equally sized frames, make one frame larger (1.5 cm in all directions) than the other and glue them to each other. Use a thin film of glue and put some weight on the wood. Deviations in the frame can influence the distance membrane-stator, so work precisely. Now the tubes can be glued on top of the larger frame (instead of between!!!). Apply some large dots two component polyurethane glue on the frame and place the tubes (which can be filled with sand) on the glue. Place a flat piece of any clean and flat material on the tube/frame and press at both sides. You can't press the tube further than the top of the frame. It is a good thing to wait a while before you press until the glue has a high viscosity. In this design, the tubes are 'floating' on a 2 mm layer of glue and you may worry about the tubes sinking in the glue with resulting deviations in the distance membrane-stator. Small pieces of 1 or 1.5 mm ('spacer') could be glued on the frame to decrease this distance and thereby preventing this, but when enough glue is used this will certainly NOT be necessary. I have successfully used this method with PVC tubes filled with sand (without 'spacer'). All the tubes were lying in virtually the same plane. This can be checked by putting a piece of sewing cotton on the tubes. Before gluing: clean the PVC with toluene (thinner) to remove the wax left during its production and make sure that there aren't any extensions at the edges. After the tubes are glued: put 'silicone kit ' between the tubes and by using a small lath, make a gradual transition between the two frames. This will diminish sound diffraction.
Some basic calculations concerning the wires: Diameter of the insulated copper wire: 1.77 mm Diameter of the nail: 2.0 mm Distance (centre to centre) between the nails: 1 + 1.77 + 2.0 + 1.77 + 1 = 7.54 mm (all wires exactly parallel) (You don't have to place the nails with such a high precision) 'Number of wires' (n) when the membrane is driven across a length of 160 mm: 160 = n * 1.77 + (n-1) * 2.0 ---Ãƒn=42.97 = 43 wires Percentage of open space of the stator (excluding the PVC tubes): 1- (n * 1.77/160) * 100% = 52% Tensioning and positioning of the wires: You have to pull at the wire very strongly to obtain visual good results. In my first ESL, I pulled not strong enough, because of the obvious risk of bending the frame, leaving some slightly coiled wires. I used thick MDF (2 * 18 mm) stators and noted only very small bending when applying a large tension. In the ESL two stator frames are mounted in opposite direction to each other, so the tension forces will cancel. The wires are wound around 2 mm thick nails. You have to drill holes of 2 mm first. When you feel an increasing resistance while hitting the nails in the holes: stop! Drill a bit further, because the frame, otherwise the frame can split. It is possible to have a split frame that looks good at first glance! (but it isn't) The wire can be glued with the tubes with Bison 'one for all'. Clean the wire by putting it in an ammonia solution during a night. Wind the wire around a large round object such as a waste paper basket (you can dry the wire now with a clean towel). If you put the wire on the frame directly from the strand you have bought, you will end up with a horrible mess of wire. After the wire is on the frame, put tape across the wires and part of the tubes, except where you want to glue the wires with the tubes. If the wooden frame is slightly curved because of the tension of the wires, place some weight one it before you glue the wires!!! Put glue between the wires (with a small stick) and remove the excess by a small lath moving across the wires leaving the excess on the tape. Remove the tape after the glue has set. Because of the lath, the glue will not protrude above the wire. If the glue protrudes very slightly (0.1 mm or less): don't try to remove it! It is not necessary and you may weaken the binding. Positioning the wires: mark the position of the each wire with a black pencil on millimetre paper (black bars) and glue this paper on a small lath. Place this lath on the tensioned wires and put tape on the wire while correcting the position according to the black bars on the paper. Although making this will take some time, your ESL will look much better when the wires are exactly running parallel to each other.
Making the membrane electro-conductive: Because the membrane is not driven at the outer 1.5 cm from the frame, you don't have to make the membrane conductive at that area. This has the advantages that possible leakage of current from the membrane to the stator is diminished and that the frame capacitance (see article of Neil S. McKean), becomes smaller. The area which has to be made conductive, can be marked with low adhesive painters tape. Don't forget to make conducting traces to the corners in order to connect the membrane with the high voltage supply. I have used graphite and that worked well. I also used detergent (in my case 'Driehoek' liquid soap), with a resulting resistance of around 30 megohm. Applying is easy, but the first time, some places were covered less by the soap (up to a tenfold reduction in conductance), resulting in serious loss of voltage across the membrane. The second time I simply added a second layer of soap, applied in the other direction. You can check the resistance by simply measuring it with a suitable device, but it is also possible to visually check the conductive layer. You can clearly see the layer of soap when light hits the membrane with the right angle of incidence. I used a dot of cotton, but I think that using a brush is just as good. (better?). Cotton releases some fibres, especially at the beginning, so first apply some soap to another surface.
Gluing the membrane with the stator The membrane can be glued with the stators (perspex glued on the wood) using Bison 'one for all'. This glue can easily be removed which is a major advantage in case of dismounting. Also, its flexibility may help to dampen some unwanted vibrations of the membrane.
When mounting the ESL element in a frame, a construction of large PVC tubes can be used to diminish diffraction. This has been done by E.Fikier and his design can be found in the Dutch magazine Hi-fi Luidsprekers, deel 11 (Elektuur). The hybrid ESL he designed was supported by three PVC tubes in which the third tube was placed behind the ESL in order to obtain a constant divergence of backwards radiated sound. This improves the interaction of the ESL with the listening room. Of course, this set up is also useful in case of full range ESLs. The tube behind the ESL should have the same diameter as the width of the membrane and has to be placed around 20 cm distance from the membrane. You may experiment with this distance or even make a tube which is moveable. I used black PVC tubes; they are much cheaper than the grey variant.
Checking the high voltage supply: I successfully used a high voltage probe (Dynatek) to check the high voltage supply and the voltage on the membrane. I measured 5 kV at the supply as well on several spots on the membrane. In case of the ESL 175, these are correct values. There is no need to buy such a probe when the ESL works correctly.
Coming to the end. I have finished the two ESL elements and they seem to work good. I still have to perform some final checks and I also have to finish the 'enclosure' made of three PVC tubes (see above). After all, I think that building an ESL yourself is not difficult. The thing you need is a good design and with simple and reproducible construction features. Also, the right choice of glue is important. You have to work with a certain degree of precision. Although an ESL is not largely affected by deviations, you also want an ESL that looks good. An attractive feature of building an ESL is that you don't need much tools and that the whole ESL element is quite two dimensional, thus easy. I hope I have given some useful information to anyone building an ESL.
I have just completelty finished one ESL; the one Hans Zeeuwe also has build. It has a 12 um thick polyester membrane coated with graphite and glued within a solid 8 cm thick MDF frame. It works fine most at the time but this is what happens when the frequency is sweeped from 10 Hz to higher: At very low frequencies (< +- 30 Hz) no problems. But at around 30-40 Hz the membrane starts to resonate severly. The whole construction suffers from this incredible resonance thereby blurring the sound. Even at moderate levels, some parts of the membrane, especially in the middle (height) of the ESL, resonates and hits one of the stators. I have got two tracks on my CD which can only be played at low volumes because of this resonance. It seems to me that this is a ''normal'' membrane resonance since increasing the resistance after the HT supply to 300 MOhm as well as temporary disconnecting the HT supply doesn't have much effect. Have I got the wrong membrane tension? And if so, could I cure this by heatshrinking the membrane? Is this due to non-uniform movement of the 'fresh' membrane with added charge migration. Will it dissapear after some period of time? Maybe 12 um thickness of the membrane is not the best choice in my application. Who helps??!!
Update June 3, 1999
The resonance problem (dynamic stability problem) has been solved by replacing the graphite coated membrane by a membrane with high resistive coating (40 Mohm). I think the problem was mainly caused by charge migration on the low resistance membrane. Also, resistance was not equally distributed. It is difficult to get equal resistance when using graphite. Another disadvantage of graghite is that it you have to work very clean to avoid spots on your carpet and to avoid discharge between membrane and stators; the small graphite particles tend to go everywhere! I will not use graphite coating for these reasons again. I reject the possibility that the problem was caused by a too low membrane tension since this would affect the static stability too. I am working on a second set of ESLs which I want to present within one month. It's construction resembles the first set, but it contains some small improvements. Also, special attention will be paid to its appearance.
MartinJan Dijkstra, The Netherlands