# About the audible noise of capacitor

All dielectric materials deform under the stress of an electric field, it's called "electrostrictive effect". And some dielectrics exhibit an additional piezoelectric effect, such ceramic capacitors, but the dominant mechanism is piezoelectric effect. I want to know if plastic film capacitors have this phenomenon? And has anyone encountered this?

• Electrostrictive effect ... yes you can sometimes hear the film capacitors in a loudspeaker crossover "sing" quietly when driven with high signal levels (and the loudspeaker itself is replaced by a dummy load, otherwise it would be at least 80dB louder!) – user_1818839 Nov 4 '14 at 11:22
• All capacitors must show this effect. But you do not need the electrostrictive effect to produce forces inside a capacitor. When loading a capacitor, the carriers inside the electrodes result in a force on the electrodes regardless which dielectric is used (even with air). But I do not know, which Voltage has to be applied to result in audible movements. A film capacitor will especially deform, if it is not fully embedded in resin or oil, so the electrodes have some degree of freedom to move. – Ariser Nov 4 '14 at 12:25
• I think engineers in power supply design filed may be familiar with this. There are articles about ceramic capacitor's noise, but no docs about that of film capacitor. – diverger Nov 4 '14 at 12:59
• @BrianDrummond, I'm not an audio guy, but are you sure it's not the inductors that are singing in the cross over network? (perhaps an active network with no inductors?) – George Herold Nov 4 '14 at 14:07
• In the specific instance I mention, absolutely sure. A Spice simulation on the shiny new departmental VAX 11-750 suggested one capacitor in the BBC LS3/5A crossover might fail at a specific frequency and voltage. The test eventually took two Quad-405 power amps, bridged together, under a fan, driving only the capacitor, but eventually it did fail... (An active network would operate at low level anyway) – user_1818839 Nov 4 '14 at 15:02

When subjected to an electric field, Dielectrics expands in the direction of the field and contract in a direction transverse to the field. The strain varies roughly with the square of the electric field. Most materials have electrostrictive strains on the order of $1*10^{-7}$ for Electric field strengths of $1 \frac{MV}{m}$. In Perovskite oxide materials the strain is $1* 10^{-3}$ at the same field strengths. Data comes from *(1) which seems to be highly cited.

While $1 \frac{MV}{m}$ seems to be unreasonable, a simple plate spacing of 1 um and a drive of 1 volt can easily meet this Field. So a normal capacitor with a 100 nm gap and 10 Volts being driven can easily exceed that number by 100 X.

For a capacitor that is say 1 mm in a given dimension a $1 \frac{MV}{m}$ Field would cause expansion by about 1 angstrom (0.1 nm) and probably more than 100 X that.

How much of that couples into an audible sound is hard to tell, but US patent # 7689390 B2 uses 100's of nm thick BST ($Ba_{0.7}Sr_{0.3}TiO_3$) layers to couple electrical and acoustic modes for a capacitor.

PTFE (Teflon) has high electro-restrictive coefficients. *(2)

So it does look possible.

(1) "Cracking in ceramic actuators caused by electrostriction" W Yang, Z Suo - Journal of the Mechanics and Physics of Solids, 1994 - Elsevier"

(2) Schwodiauer, R.; Neugschwandtner, G.; Bauer-Gogonea, S.; Bauer, S.; Heitz, J.; Bauerle, D., "Dielectric and electret properties of novel Teflon PTFE and PTFE-like polymers," Electrets, 1999. ISE 10. Proceedings. 10th International Symposium on , vol., no., pp.313,316, 1999

• Thanks, can you give some information on PET and PP, which are more common in practical applications. – diverger Nov 7 '14 at 6:01