The motor-run capacitors are far from identical to audio capacitors, or even to each other. Since this often comes up I'll toss this in this explanation, and requote as needed.
This is from my notes when I was investigating whether a motor run capacitor using dry polypropylene film was a better solution for crossovers than the conventional polypropylene film sold for audio use. The claims of superiority for the motor-run capacitors over the conventional audio equivalent is, I concluded, yet another form of audio fetishism where giant capacitors of dubious internal construction are deemed to be superior based upon the appearance. The claims are not based in reality, but is oft repeated enough to have the force of audio dogma.
I am setting forth sufficient detail to permit all of this to be independently researched and verified. I'm sure the golden ears will chime in about "
to my ears".
The standards for motor-run capacitors are, in part, governed by loose standards permitting a
2% to 5% variation in capacitance with temperature, over and above the capcitor's nominal tolerance. Some capacitors may have greater tolerance.
The motor-run capacitors are also designed for AC use, which, being 50 or 60 Hz, is far, far lower than audio frequencies. The properties are not the same as frequency increases, and some of the datasheets provide quite different values at 1,000 Hz than at the normal motor frequency of 50 or 60 Hz.
(1) Film versus Electrolytic
The first bifurcation between the motor-run capacitor types occurs between
film versus electrolytic.
Modern motor-run capacitor are now commonly fabricated as NPE as a
cost-savings measure. The lower lifespan of the equipment in which such capacitors are installed, and the associated cost-savings for less-expensive components, therefore benefits from NPE use by way of the lower transfer price in the retail channel, which usually wins over the potentially increased maintenance costs down the road. The cost to replace a failed, failing, or merely old NPE capacitor during routine maintenance is considered to be lower than the cost to purchase film at both time of manufacturing and for ongoing maintenance.
One must therefore be careful to select the proper type of capacitor. I have seen photos of crossovers where the builder crows about the sonic benefits when the capacitors are identifiable as NPE by way of the markings on the side.
The typical film is polypropylene, but this is not guaranteed and all film types are not equivalent, particularly when oil is used as an impregnant. Again, reading of manufacturer specifications is required.
(2) Dry Film versus Wet (Oil) Film
The second bifurcation is
dry film versus wet film. Each is further discussed.
(2a) Variations in Dry Film
The film may be polyester, polypropylene (PP), or some other plastic. This matters.
Polyester (polyethylene terephthalate aka PET) — Mylar is the Dupont and Teijin Film Enterprises trademarked name for polyester — has drawbacks:
(a) Profound dielectric absorption which performs signal averaging. The significant difference in dielectric absorption must not be ignored. PET is 0.5% while polypropylene 0.05%, i.e. a factor of ten worse. This accounts for the sonic differences. The reason, in brief, is migration of electrons into the dielectric which are then released over time. Think of this as signal averaging. It is a significant issue in low-noise opamps because the capacitors can integrate over time. Anyone who pushes PTFE (tradename Teflon) for interstage coupling capacitors claiming it has cleaner sound, and a dielectric absoprtion of 0.01% (that's 1/5th of PP and 1/50th of PET) might want to think about using PET instead of PP.
(b) PET is hygroscopic, and any retained moisture causes depolymerization with heat, and that heat arises from ESR which is more significant with constantly changing AC signals. .
(c) Positive temperature coefficient which cause capacitance to rise with temperature. This is not an issue for motor-run applications, but might be for audio. Particularly given that the ESR may give rise to ohmic heating.
(d) PET has high affinity for oil, able to absorb thirty times its weight, which is three times better than cotton which can absorb roughly ten times its weight. This is why PET is used for petroleum oil spills in oceans and waterways. This has ramifications for wet capacitors, as below set forth.
(e) Dissipation Factor is different. For the capacitor I spot-checked, PET is less than 7 x 10-3 while polyproplyene is less than 5 x 10-4., approximately ten times lower.
The advantage of PET is the 50% larger dielectric constant (ε = 3.3 versus 2.2), which increases capacitance 50% per unit volume.
While the cost is trivial per capacitor, this is a high-volume and very competitive business, so tiny differentials to matter.
(2b) Variations in Wet Film
A wet film capacitor immerses the internals, typically metalized film, uses oil to:
(a) Increase voltage rating, as the oil is non-conductive and displaces air bubbles which would otherwise arc at voltages below the rated limits (both AC and DC)
(b) Reduce corrosion of the metalization layer and contacts, which may be heterogenous as a cost-savings measure, with metals less expensive than aluminum commonly used
(c) Improve heat dissipation for the ohmic heating produced by the ESR.
The film layer, however, may over time absorb the oil and swell.
While the changing the capacitor geometry may slightly alter the capacitance the more important issue is that the film or its metalization layer may eventually crack from the expansion or the contact points may somewhat detach, reducing conductivity and increasing ESR and thus ohmic heating. This is a known reliability issue with wet-film capacitors, and selection of oil is crucial. Not all manufacturers may properly perform such selection, given the commodity nature of, and low price-point for, such capacitors, or may consider it to be a non-critical issue from the perspective of lifespan which, again, is as short as the equipment into which it is installed.
Swelling is negligible with vegetable oils but petroleum oils, both natural and synthetic or reprocessed, easily swell polypropylene film. The oil must be degassed to prevent the formation of air pockets.
The swelling of the film causes weakening of the end connection between the deposited electrode and the metal sprayed at the end, terminating in open circuits under severe conditions.
Again, the application of motor run is different in that the motor or compressor typically fails before the capacitor, and maintenance swapping for capacitors is considered to be a normal and routine practice.
(3) Differences in Foil vs. Metalization
The third bifurcation is based upon plate construction, i.e. foil vs. metalization, such that (a) physically distinct layers are used for the dielectric (film) and plate (metal), or (b) the metal plate is directly deposited upon the film dielectric using a plasma arc, so one layer serves both purposes.
A foil process consequently has a much thicker plate, whereas metalization (vapor deposition or sputtering) is so thin that the metalization/dielectric layer would be transparent when held up to the light, much like metalized sunglasses. The foil consequently has lower resistance than metalization, which greatly improves charge distribution for AC and pulsed DC applications. But the metalization permits smaller capacitors with less metal and thus lowers the size and cost.
To further reduce manufacturing costs, the motor-run capacitors may further compromise the properties for audio by replacing the metalization layer of pure aluminum with either zinc or a hybrid of aluminum and zinc, or even some other alloy. This matters as zinc has a conductivity of 27 relative to copper and aluminum has 61, again relative to copper, which is 2.6 times greater. (Conductivity is expressed as IACS, International Annealed Copper Standard, with copper being 100 or fully conductive.) Chinese motor-run capacitors commonly use zinc, but many of the US manufacturers are making capacitors (typically in China) using zinc. Yet another race to the bottom.
Zinc increases ESR which is not an issue for the intended application, i.e. motors, but may matter for audio. Interesting how audiophiles who pay high premiums for oxygen-free copper wriing, low-gauge air-core inductors (sometimes even flat inductors), copper-foil coupling capacitors, and silver fuses will happily use capacitors with low-conductivity zinc or alloys, and then crow about the sonic benefits.
(4) Differences in Construction
The fourth bifurcation concerns the internal attachment points, which differ between motor-run and audio capacitors.
Audio capacitors use spot-welding, which physically melts together two metals, similar or dissimilar, to form a low-resistance contact. The motor-run capacitor, however, uses other attachment techniques to reduce costs, typically a non-welded and non-soldered pressure fit, or a pressure with a conductive paste added to improve conductivity. Zinc is often used for this connection point. Higher ESR is the result, and this can be high enough to increase dissipation by ten to a hundred times. In addition, the contact point may corrode, hence the need for oil or a sealed container, which also increases ESR.
TL/DR
Best to stick with audio grade polypropylene capacitors as these are small, inexpensive, and do not have the deleterious properties above described.
