Conductive coating is one of the most important elements affecting the sound quality and reliability of electrostatic loudspeaker. For sure, there are many so called "most important" variables but this one is exceptionally decisive when it's not working properly. Coating should satisfy many conditions :
- Long-term stability
- Resistance to UV light(important for designs without protective cloth)
- As little as possible variation depending on relative humidity
- Resistance not too high, typically below approx 10^10 Ohm*sq(A notable exception is quad ESL 57).
- Resistance not too low, typically above 10^8 to limit change migration.
- Low added mass
Unfortunately it is very difficult to find a coating which satisfies all those conditions, so often compromises are not possible to avoid.
Some products have been tried with varying success. Short overview is presented below.
a) Graphite rubbing & wiping with alcohol. A lot of messy work; surface resistance typically too low. Difficult to achieve uniform conductivity. Long term stability is difficult to predict as reaction between carbon and oxygen possible under corona.
b) Staticide 2001 / 2003. This coating seemed to work more or less. Variation depending on relative humidity was high. Resistance too low; although it was possible to increase it by diluting.
c) Polycoat from MT Audio design page. Low variation depending on ambient humidity; easy to achieve uniform resistance. However surface resistance of around 10^7 Ohm*sq is somewhat below the optimum range.
d) DuPont Elvamide. A little more about that.
At first I would like to note that there are several grades of this soluble nylon : 8061, 8063, 8023R and 8066. The material is similar to what was used in Quad ESL 57(it was called Calaton according to other sources). The conductivity is acquired through water absorption. If one is about to examine the datasheet of Elvamide it becomes apparent that different grades have considerably varying moisture content. Therefore grades 8023R and 8066 would have increased conductivity compared to 8061 & 8063. According to this source Calaton was prepared from Nylon 6,6 which is believed to be more or less similar to Elvamide 8066. Therefore it is assumed that the most common grades 8061 & 8063 should have much lower conductivity due to lower moisture content compared
to Calaton.
e) Licron crystal. Update : surface conductivity has dropped below acceptable levels for one panel. The ambient humidity is at around 40%. Reasons are not clear.
Application. A spray can was drained and liquid applied by using a cotton wipe or brush. At first the thinnest possible layer was experimented with. Loss of sensitivity was observed after a few weeks. It seems that this coating has a tendency to increase it's resistance after long periods of time. Considerably thicker layer was tested later. The usage was about 7ml for panel area of 0,34m2. This equals to approx. 20ml/m2.
One special note is about how the contact between coating and EHT supply is done. Conductive self-adhesive copper tape is used. A thin line of coating is applied via brush at the point where copper meets the diaphragm. This is required to prevent corona discharge, which could "eat" the coating away at contact point(thanks Bolserst from DIYaudio forums for suggesting this).
- Long-term stability
- Resistance to UV light(important for designs without protective cloth)
- As little as possible variation depending on relative humidity
- Resistance not too high, typically below approx 10^10 Ohm*sq(A notable exception is quad ESL 57).
- Resistance not too low, typically above 10^8 to limit change migration.
- Low added mass
Unfortunately it is very difficult to find a coating which satisfies all those conditions, so often compromises are not possible to avoid.
Some products have been tried with varying success. Short overview is presented below.
a) Graphite rubbing & wiping with alcohol. A lot of messy work; surface resistance typically too low. Difficult to achieve uniform conductivity. Long term stability is difficult to predict as reaction between carbon and oxygen possible under corona.
b) Staticide 2001 / 2003. This coating seemed to work more or less. Variation depending on relative humidity was high. Resistance too low; although it was possible to increase it by diluting.
c) Polycoat from MT Audio design page. Low variation depending on ambient humidity; easy to achieve uniform resistance. However surface resistance of around 10^7 Ohm*sq is somewhat below the optimum range.
d) DuPont Elvamide. A little more about that.
At first I would like to note that there are several grades of this soluble nylon : 8061, 8063, 8023R and 8066. The material is similar to what was used in Quad ESL 57(it was called Calaton according to other sources). The conductivity is acquired through water absorption. If one is about to examine the datasheet of Elvamide it becomes apparent that different grades have considerably varying moisture content. Therefore grades 8023R and 8066 would have increased conductivity compared to 8061 & 8063. According to this source Calaton was prepared from Nylon 6,6 which is believed to be more or less similar to Elvamide 8066. Therefore it is assumed that the most common grades 8061 & 8063 should have much lower conductivity due to lower moisture content compared
to Calaton.
e) Licron crystal. Update : surface conductivity has dropped below acceptable levels for one panel. The ambient humidity is at around 40%. Reasons are not clear.
Application. A spray can was drained and liquid applied by using a cotton wipe or brush. At first the thinnest possible layer was experimented with. Loss of sensitivity was observed after a few weeks. It seems that this coating has a tendency to increase it's resistance after long periods of time. Considerably thicker layer was tested later. The usage was about 7ml for panel area of 0,34m2. This equals to approx. 20ml/m2.
One special note is about how the contact between coating and EHT supply is done. Conductive self-adhesive copper tape is used. A thin line of coating is applied via brush at the point where copper meets the diaphragm. This is required to prevent corona discharge, which could "eat" the coating away at contact point(thanks Bolserst from DIYaudio forums for suggesting this).
Measuring surface resistance of the coating.
Special probes exist for this purpose; however those are quite expensive. A very simple tool can be built at home by using a digital multimeter and some contact surface as probes. The photo of the device is shown below. Voltage source is 9V battery. It's important to make sure surface on which mylar is placed is not conductive. For example, a typical wooden tabletop might allow current to flow through it and show false figures. So either glass or plastic is recommended. Suspending membrane in air or measuring when glued to the speaker is even better. Electrodes can be made from a pair of large coins or a piece of PCB for example.
Since input resistance is known, it is easy to calculate surface resistance of the film. In the example below its approx R = 9/1.76 * 1*10^7 = 5.1*10^7 Ohms square.