The Sound of Capacitors

Capacitors are fundamental components in many electronic circuits, including audio circuits, thanks to their ability to store and release energy. Their main function is to filter, block, or couple signals, depending on the context. Despite their apparent simplicity, they can significantly influence the behavior of a circuit, especially in audio amplifiers, where every detail can make a difference in sound quality.

Many people believe that there is a distinctive sound associated with different types of capacitors. This article will explore that aspect by examining the linearity or “deviations from linearity” of the various types of capacitors used in audio applications.

In addition to linearity, other technical aspects that influence the behavior and sonic quality of capacitors will be examined, such as ESR (Equivalent Series Resistance), ESL (Equivalent Series Inductance), dissipation factor, and leakage. These parameters not only determine the performance of capacitors in audio circuits but can also contribute to what many people perceive as the “sound” of capacitors.

To write this article, I wanted not only to draw on my own knowledge, but also to save from oblivion an important contribution: the article “The Sound of Capacitors” by Steve Bench, published around the mid 1990s. The original article has been offline for decades and now only survives on a few mirror sites that could disappear at any moment. It seemed important to me to preserve and integrate the information from that work in an even more exhaustive context.

In addition, there was a collaboration with “Pier Aisa – Elettronici Entusiasti” – “Pier Aisa – YouTube Channel“, who produced a video where he explains these same concepts, trying to replicate Steve Bench’s original measurements with Lissajous figures on capacitors, just like about 30 years ago.

Link to Pier Aisa’s video

I would like to thank Manuel for the photo of the tantalum capacitors.

Reactive Components: Fundamentals and Their Importance in Audio

In the world of electronics, we often talk about reactive components, which include capacitors and inductors, as well as transformers. Reactive components are characterized by their ability to store electric or magnetic energy, generating a reactance that can influence circuit behavior depending on signal frequency. Audiophiles are well aware of how crucial transformer quality is in tube amplifiers, since their performance can significantly affect the final sound. In the same way, the quality of the capacitors used in these circuits is equally important. Their characteristics can be perceived when listening, contributing to the reproduction of a richer and more detailed sound, or simply a “different” sound. In fact, the final sonic result of an amplifier depends greatly on reactive elements, much more than on the type of tubes used. In this article, however, I will focus exclusively on capacitors, exploring their technical characteristics and their impact on tube audio amplification.

What a Capacitor Is and Its Main Types

A capacitor is an electronic component capable of storing electrical energy in the form of an electrostatic field between two conductors (plates) separated by an insulating material called a dielectric. It is used in a great many circuits, from filters to oscillators, from power supplies to audio amplifiers. Its ability to store and release energy in a controlled way makes it essential for signal transfer and voltage stabilization.

Understanding the different types of capacitors is fundamental for optimizing the performance of electronic circuits, particularly in audio. Each type of capacitor has unique characteristics that affect not only the capacitance value, but also its behavior at various frequencies, stability, long term reliability, and audible sonic character. There are therefore several types of capacitors, each with specific features that determine their use in different applications:

  • Electrolytic capacitors: used for high capacitance values at low frequencies, they have a defined polarity.
  • Tantalum capacitors: similar to electrolytics but more compact.
  • Film capacitors: characterized by stability and low loss, used in audio and precision circuits.
  • Ceramic capacitors: very common, they offer a wide range of values and can be used in high frequency circuits.
  • Silver mica capacitors: used in high frequency circuits for their low loss and high precision.
  • Supercapacitors: used to store large amounts of energy, although they have a low operating voltage.
  • Glass capacitors: used where very high stability and precision are required in RF and high voltage applications, but less common due to their cost (they are not of interest here, I mention them only for completeness).
  • Paper and oil capacitors: once widely used, modern models differ significantly from the old ones, since the oil used in the past is now banned for environmental reasons.

For each type of capacitor, I will show some identifying photos and measurements that may be of interest, in addition to the images provided by Steve Bench. Steve explained how he obtained these Lissajous figures by measuring different types of components and capturing the results simply by pointing a camera at the oscilloscope. The value of each capacitor was kept constant at 0.1 µF. The signal level was kept constant at about 70 volts RMS at 600 Hz through the capacitors, with a signal current of around 26 mA. This current level is probably higher than what you would normally expect, but it serves to show the results more clearly. Several types of capacitors were used in the experiment, including paper and oil capacitors, polycarbonate film, polyester film, polystyrene film, polypropylene, 100 volt and 1000 volt ceramic capacitors, and silver mica.

Let us see how the symbols of the different types of capacitors are represented in these beautiful vintage illustrations…

ESR (Equivalent Series Resistance) and ESL (Equivalent Series Inductance)

ESR is the internal resistance of a capacitor. Although ideally a capacitor should behave as a purely capacitive device, in reality all capacitors exhibit a small internal resistance. ESR affects capacitor losses, particularly at high frequencies, and can contribute to heating of the component during operation. Low ESR is desirable for better performance, especially in high current, high frequency applications such as amplifier power supplies.

In audio applications, ESR is particularly relevant in electrolytic capacitors, because high ESR can negatively affect the capacitor’s ability to supply very fast current peaks, which is essential for the correct operation of power supply stages. In addition, high ESR can limit the capacitor’s ability to shunt unwanted frequencies to ground, compromise signal quality, and reduce the effectiveness of filters. This is especially important in amplifiers, where frequency response and signal dynamics are fundamental for optimal audio performance.

ESL, on the other hand, represents the inductance created in the capacitor due to the geometry of its conductors and connections. Even though a capacitor is designed to have only capacitance, its physical construction introduces a certain amount of inductance. This is particularly problematic in high frequency applications, where ESL can cause less than ideal behavior, limiting the effectiveness of the capacitor as a filter.

Leakage Current

Leakage in capacitors is a phenomenon that relates to energy losses in the dielectric itself, which manifests as a reduction in charge over time or with rising temperature. These losses are normal and do not necessarily indicate a defective capacitor. They are instead a consequence of the inherent inefficiency of electrostatic materials, which in some contexts can create problems.

Relationship Between Dissipation Factor and Frequency

The dissipation factor (D) is a crucial parameter that describes energy losses in a capacitor. It represents the ratio between the loss resistance and the capacitive reactance. It is calculated as follows:

D = R_loss / X_C

where R_loss is the effective resistance representing losses (including those due to the dielectric) and X_C is the capacitive reactance of the capacitor. It is important to note that the dissipation factor is influenced by the frequency of the signal applied to the capacitor. As frequency increases, the capacitive reactance (X_C) decreases, since it is given by the formula:

X_C = 1 / (2 * pi * f * C)

As a result, a decrease in X_C leads to an increase in the dissipation factor, if the loss resistance remains constant. In audio applications, the dissipation factor (D) plays a key role in film capacitors. A low and stable D with respect to frequency is synonymous with lower losses, greater linearity, and superior sound quality. When the dissipation factor is high or varies significantly with frequency, it can introduce distortion and degrade the audio signal, compromising the listening experience. In particular, low quality capacitors tend to show an increase in dissipation factor as frequency rises. For this reason, choosing capacitors with an optimal and stable dissipation factor is essential in high fidelity audio applications, where every circuit element can significantly influence the final result.

Relationship Between Capacitance and Frequency

It is important to point out that although it is possible to measure a capacitor’s capacitance as frequency varies, these measurements usually show only minimal variation and provide little meaningful information about the real behavior of the component. It is much more useful to focus on parameters such as ESR and dissipation factor, which provide far more relevant indications of actual capacitor performance. Capacitors with high ESL may cease to function correctly at lower frequencies than those with lower ESL. This phenomenon becomes evident when measuring capacitance as a function of frequency, because at a certain point the instrument will indicate a negative capacitance, showing that the capacitor has stopped acting as such and has effectively turned into a pure inductance. However, the usefulness of this type of measurement is limited to this aspect alone. Small variations in capacitance with frequency are insignificant in themselves and instead reflect other parameters like D, ESR, and ESL, which have a much more important impact on the sonic behavior of the component.

So it is essential not to pay attention to certain gurus who show only capacitance versus frequency measurements. These individuals pretend to be educating you, but in reality they are only trying to create more mystery and disinformation, once again pushing the wrong idea that what seems to measure better on instruments must sound worse and vice versa. Furthermore, the poor high frequency behavior of certain capacitors, particularly electrolytics, which have high ESL and ESR, can be mitigated with tricks and expedients that I will discuss later in the article. The analysis of ESL and ESR, in particular, becomes essential in high frequency applications, since high parasitic inductance can compromise the effectiveness of the capacitor in filtering or coupling signals, making capacitor selection critical in the circuit.

Dielectric Linearity

In capacitors, dielectric linearity is a crucial factor for performance. In a vacuum, the dielectric constant is equal to 1, which means that the capacitor has highly predictable and stable behavior. Air has a dielectric constant very close to 1, but slightly higher. However, inserting materials such as paper, plastic, mica, or oxides between the capacitor plates increases the capacitance, because these materials have a dielectric constant greater than 1.

Although this higher capacitance is undoubtedly useful, many of these dielectric materials can introduce nonlinearities that cause the behavior of the capacitor to vary as a function of the applied voltage and the current flowing through it. Even though these materials are not linear, we are forced to use them. In fact, air capacitors with the same capacitance values would reach dimensions comparable to a football field.

Many people are already familiar with the nonlinearities in magnetic materials, described by the B-H curve, whose slope is directly related to the permeability of the material. In a similar way, dielectric materials also show comparable behavior, represented by the D-E curve, where the slope is associated with the capacitance value. This highlights how the electrostatic properties of materials can influence their behavior in a circuit. Such effects can degrade performance, particularly in sensitive audio applications where signal fidelity is a priority. The use of nonlinear dielectrics can alter frequency response and compromise sound quality.

How to Obtain the D-E Curve

Once again, referring to the more familiar “magnetic” analogy, we can obtain a B-H curve (actually a PSI versus I curve) by observing the E-M relationship: E = L dI/dt

If you sample the current flowing through an inductive circuit (on the x axis) and integrate the voltage across the inductor (on the y axis), you will obtain a curve proportional to the B-H curve. For a capacitor, the operating equation is: I = C dV/dt

Consequently, if you sample the voltage and integrate the current, you will obtain a curve proportional to the D-E curve of the component. Surprisingly, real components show a CURVE rather than a straight line. They may also display hysteresis, just like inductors or transformers. This phenomenon introduces subtle forms of distortion and nonlinearity that can compromise the accuracy of music reproduction.


Electrolytic Capacitors

In tube amplifiers, electrolytic capacitors are commonly used in the power supply section and as cathode bypass capacitors for the tubes. A crucial parameter in these applications is ESR (Equivalent Series Resistance), since high ESR in a filter or decoupling capacitor, or in a cathode bypass, can negatively impact the sound, making it muddy or dull. However, thanks to the low currents typical of tube circuits, it is often possible to correct the shortcomings of electrolytics by adding a good quality film capacitor in parallel. Values between 220nF and 1uF are usually sufficient to solve these issues.

In circuits that require higher current peaks, such as solid state power supplies, several electrolytics can be placed in parallel to reduce parasitic resistance (ESR), which is shared among the various capacitors. Even in this case, adding a film capacitor in parallel remains a useful solution. However, care must be taken not to overdo it. There have been cases where entire PCBs were filled with dozens of small electrolytics in parallel, which can become ridiculous and counterproductive. In fact, the parasitic resistance of the tracks connecting the capacitors may introduce new losses, defeating the original purpose. It is also important to note that large value electrolytics tend to have higher ESL (Equivalent Series Inductance) than smaller ones, which is why balanced design is essential for good sonic performance.

DC on electrolytic capacitor curves. In various discussions, it has been pointed out that these components work more effectively when biased with a direct voltage (DC bias). To explore this aspect, I decided to plot some curves considering different DC bias values. The typical nonlinearities of electrolytics, which mainly appear as hysteresis, can be mitigated with proper bias, reducing the extent of hysteresis. However, even with adequate bias, electrolytic capacitors still show a “curved” characteristic that reflects their intrinsic nonlinearity. Among them, the worst are non polarized electrolytics, which show even more pronounced hysteresis than their polarized counterparts.


Tantalum Capacitors

Tantalum capacitors are a solid state version of electrolytics, since they use a tantalum oxide dielectric instead of a liquid in contact with the dielectric itself. This means they do not wear out and, for the same capacitance value, they are much less bulky than traditional electrolytics. Although they are known for their stability and reliability, especially in critical environments, they exhibit quite unfavorable hysteresis curves, which implies poor linearity. In audio applications, it is fair to say that these capacitors are not suitable and that their sound quality is very disappointing. Their ideal field of application is inside computers, smartphones, and digital electronics in general, where they perform excellently. However, they must be kept away from audio applications. Below, I have included the Lissajous figures measured by Steve Bench some 30 years ago, which leave no room for doubt, even though they tend to behave better when biased at a higher voltage.


Film Capacitors

Film capacitors are valued for their stability and low loss. Made with dielectric materials such as polypropylene or polyester, they are used in audio and precision circuits where it is essential to maintain high signal quality. These capacitors offer excellent performance in terms of linearity and tolerance, making them ideal for applications where distortion must be kept to a minimum. Compared to electrolytics, film capacitors do not experience wear in audio applications. Wear in these capacitors is usually seen only in very demanding applications, such as motor power factor correction or situations where the component must filter very high voltage spikes.

In tube amplifiers, this type of capacitor is the most commonly used for stage coupling, and the most important parameter to consider is the dissipation factor. As already mentioned in this article, to achieve pleasant sound the dissipation factor must remain as constant as possible with changing frequency. Lower dissipation factors correspond to a brighter, more detailed sound, while higher dissipation factors lead to a warmer sound with fewer details in the high range.

At the two extremes of the scale, we have polypropylene, which is the material with the lowest D, followed by polyester, while at the opposite extreme we find paper and oil capacitors, which show the highest D values. It should be noted that capacitors whose D increases with frequency usually do not produce high quality sound and, if placed in parallel with electrolytics, do not provide any improvement compared to the lone electrolytic. It is also important to point out that many industrial capacitors or low cost NOS parts perform very well in audio applications, in contrast to so called “audiophile” capacitors with decorative outer casings but no real superior characteristics.

Finally, the dissipation factor versus frequency is not listed in any datasheet and, to find it, you need an LCR bridge capable of measuring at least at 100 Hz, 1 kHz, and 10 kHz.

I will now show three videos in which I measure the dissipation factor (D) of several capacitors. I will start with a polypropylene capacitor, then a good polyester, and finally a poor quality capacitor. These measurements allow us to compare the performance of each type of capacitor.

  • A high quality polypropylene capacitor, although an industrial part, is perfectly comparable to a Mundorf Supreme Classic in terms of performance. The dissipation factor (D) remains stable around 0.0001 across the entire frequency range, demonstrating excellent behavior.
  • Now we move on to a very good polyester capacitor, which shows a dissipation factor (D) of about 0.0009 at 10 kHz. This value suggests that it will probably have a slightly warmer sound than polypropylene, while still maintaining good characteristics.
  • Now we move on to a rather poor quality polyester capacitor, with a dissipation factor (D) starting at 0.0018 at 100 Hz and reaching 0.09 at 10 kHz. Although these values are not ideal, I have seen even worse performance, especially in “audiophile” capacitors of Chinese manufacture where quality is often questionable. The key point is not only whether the dissipation factor (D) is high or low, but how constant it remains as frequency changes. In this case, D increases greatly with frequency, which makes this green capacitor poorly suited for audio. When there is a pronounced increase in D between 100 Hz and 10 kHz, sound quality suffers significantly. The larger this jump, the worse the perceived sound.

Now let us look at the hysteresis curves. In general, these curves show very low distortion levels and are of little relevance for most audio applications. From the standpoint of dielectric distortion, polycarbonate stands out for its interesting characteristics. It is possible to adjust its performance by combining it in series or in parallel with other types of capacitors. Polystyrene, on the other hand, sits halfway between paper and oil capacitors and polycarbonate, and offers performance that is almost comparable to the former. By examining the polypropylene curves carefully, you can see a slight curvature, which makes it similar to polystyrene in terms of linearity, while preserving its own unique “signature.” In general, however, this aspect can be ignored for film capacitors and I can say that the only significant parameter for audio is the dissipation factor.

Today, the two most widespread types of film capacitors are polyester (MKT) and polypropylene (MKP), while polystyrene and polycarbonate were more common in the past and are often found as NOS (New Old Stock). It is also worth mentioning, for completeness, capacitors made with cellulose acetate (MKL), which are almost unknown today but do exist.

Film capacitors manufactured from the late 1970s onward should be built using non inductive techniques, which means they have low ESL (parasitic inductance). This allows them to operate well even at relatively high frequencies, well beyond the audio range.

However, there are exceptions represented by older NOS capacitors whose rolled plastic film is clearly visible, as shown in the photos above. These capacitors, in addition to being sensitive to moisture, also display inductive behavior. Although they usually do not cause problems in the signal path, if used as compensation in a negative feedback loop they may cause instability and generate RF spurious signals (at least that has happened to me).


Ceramic Capacitors

Ceramic capacitors are among the most common parts on the market and offer a wide range of capacitance values. Thanks to their small size, they can operate effectively in high frequency circuits, which makes them ideal for telecommunications and signal circuits. They also have the lowest ESR and ESL of all, which ensures greater stability of capacitance with frequency. These qualities make them excellent wherever very high frequencies and radio signals must be handled.

However, despite these technical advantages, ceramic capacitors are known for their poor sound quality. This is where some snake oil guru might step in and say: “You see, measurements do not matter, because what looks good on instruments does not always sound good.” This is the right moment to debunk that claim. Ceramic capacitors are actually among the worst in terms of hysteresis, producing unacceptable distortion. The nonlinearities associated with these capacitors are obvious and severe enough to be audibly unpleasant. Their value changes with temperature and they are microphonic as well.

Their characteristics change dramatically with frequency. Some people claim that these peculiarities can be exploited to add a “crunch” effect to guitar amplifiers (amplifiers which are deliberately designed to distort) by using capacitors with different rated voltages (as long as they are adequate for the amplifier). During measurements, two types of ceramic capacitors were examined: a 100 volt Z5U and a 1000 volt Z5U part. In addition, there is a very common monolithic bypass capacitor used in digital logic.

In the last image, in addition to the obvious nonlinearities at the extremes, you can see another “kink” around the zero crossing point, which contributes to a “delightful” sonic character. There is also another concerning aspect: the performance of these capacitors varies markedly with frequency. So we present data at 600 Hz and 100 Hz, highlighting the hysteresis, “saturation,” and strange kinks that characterize their response.

In audio amplifiers, using ceramic capacitors for signal coupling is something that must absolutely be avoided. However, they can be used in specific situations where you want to solve problems related to ultra high frequencies, such as instability or oscillations that occur far outside the audible range. In these cases, the “audio” quality of the capacitor has no impact on overall circuit performance. Ceramic capacitors can also be used for feedback compensation, as long as the resonance to be attenuated is always at very high frequencies. For frequencies lower than RF, it is better to choose silver mica or film capacitors, which offer superior audio performance.


Silver Mica Capacitors

Silver mica capacitors are known for their low loss and high precision, which makes them ideal for high frequency circuits. These capacitors use mica as the dielectric, which provides excellent thermal and mechanical stability. Thanks to these characteristics, they are often used in radio circuits and RF applications where absolute stability is essential.

It is important to note that mica capacitors are generally available only in very low capacitance values, while higher values can be extremely expensive. They are commonly used in tuned circuits and RF applications and may occasionally be employed in certain audio applications where using a ceramic capacitor could have compromised sound quality. Mica exhibits very mild hysteresis.


Supercapacitors

Supercapacitors represent a relatively new technology in the energy storage landscape, developed thanks to nanotechnologies. Unlike traditional capacitors, supercapacitors can store large amounts of energy in a compact form factor and handle rapid charge and discharge cycles. These features make them ideal for applications requiring fast energy delivery, such as electric vehicles and renewable energy systems.

Although their operating voltage is generally low, their very high capacitance can be extremely useful in contexts like tube amplifiers. In particular, they can be used for power supply filtering for directly heated tube filaments, such as the 26 tube, which is especially sensitive to residual hum on the filament and requires a supply of 1.5 volts and 1 ampere. Using a simple CRC cell and supercapacitors, it is possible to provide the correct power, also obtaining a soft start effect due to the charging time of the capacitor placed after the resistor, all without any active components. In the photo below you can see two 5 farad, 5 volt capacitors (which can also be used for a 300B filament, for example).


Paper and Oil Capacitors

Paper and oil capacitors were very popular in the past for various electronic applications, primarily because of their high resistance to electrical stress. They can be considered almost self healing. If an internal discharge would permanently damage other types of capacitors, in paper and oil capacitors this is not such a big issue. Thanks to the oil as dielectric, it cannot be punctured in the same way as a solid. While they are not indestructible, this characteristic made them ideal for applications like power factor correction or motor start capacitors in electric motors.

These must not be confused with the paper capacitors found in vintage radios, which contain no oil at all but only paper, sealed in a glass tube with bitumen. Those were well known for being extremely unreliable and vulnerable to moisture. At the time, before the introduction of plastic film capacitors, paper and oil capacitors were also widely used in precision electronics and military equipment, where reliability was essential. For example, in aircraft radio communications, it was unacceptable for a rainy day to impair communication.

These capacitors are made of paper sheets separating the two plates, immersed in insulating oil inside a hermetically sealed metal case. The oil not only enhances the dielectric properties of the paper but also acts as a cooling medium and prevents moisture from entering the capacitor, providing longer life and stability over time.

With the evolution of technology and the introduction of plastic film capacitors, along with increasingly strict regulations and growing environmental concerns, the production of paper and oil capacitors changed radically until they disappeared completely from modern electronics. The oil used in the past, which was largely responsible for their qualities, was banned because of its potential harmful effects. Only in the audiophile world, to satisfy enthusiasts, were capacitors reintroduced under the name “paper and oil,” which often have little in common with the originals, since they are made differently and with different oils. In some cases, especially in Chinese paper and oil capacitors, it was discovered that simple vegetable oil was used, which decomposes over time and eventually causes short circuits. I have also seen Jensen capacitors mounted in Audio Note amplifiers where degradation caused severe leakage, driving the 300B tubes into overload with red hot plates because the grid had gone positive.

We should not forget the so called “Film Oil” capacitors, which are actually simply polypropylene capacitors with oil added inside. Although their construction is similar to common polypropylene capacitors, they are nowhere near the characteristics of real paper and oil capacitors. At least if the oil inside them degrades, it will not cause a short circuit. As a result, modern paper and oil capacitors have different characteristics compared to their predecessors.

Let us now examine the distortion behavior of real paper and oil capacitors. It is surprisingly clean, probably the best of all types. This may explain why some enthusiasts prefer the sound of paper and oil capacitors. Although there are not many documented measurements, these capacitors tend to show high linearity and stability even as frequency changes. Below I show a set of paper and oil capacitors together with the same type measured at 100 Hz, to highlight the stability of their performance.

Let us now examine the dissipation factor at different frequencies for a military grade paper and oil capacitor. Slight variations can be seen, but the average value remains between 0.05 and 0.07. This means that the capacitor offers good sound, although compared to polypropylene it has less detail and is less bright in the high frequencies, resulting in the classic “warm sound.” It is important not to generalize, though. If the device already tends towards a “dark” sound, using a paper and oil capacitor may make things worse. In that case it would be better to choose a polypropylene capacitor to “brighten” the sound. On the other hand, if the device is too bright, paper and oil capacitors may help balance the sound, adding warmth.

NOTE: I am not completely sure that the capacitor in the video is paper and oil. It might be a paper and oil capacitor of exceptional quality or it may be made of some other material unknown to me…

I would like to add the dissipation factor measurement of a Vitamin Q capacitor, which definitely is a paper and oil type. In this case, you can see that D increases with frequency, showing a sudden jump between 8 and 10 kHz. This suggests that the capacitor may hide detail in the high range and might appeal to those who want to reduce sibilant sounds…

Not everything that glitters is gold, and it is not always oil that sizzles… sometimes it is the tubes that are frying…

As with everything else, there are no flawless solutions, and unfortunately paper and oil capacitors also have their downsides. I have already mentioned Chinese capacitors filled with frying oil that ended up shorting out, but deterioration also affects genuine vintage paper and oil capacitors. We do not know for sure how they behaved when new because we would need a time machine to test a fresh one. However, I can say that NOS components found today are not always in perfect condition. On average, almost all paper and oil capacitors show some internal leakage, in other words, they have a certain amount of leakage current.

For comparison, a polypropylene capacitor measured with a 500 volt gigaohmmeter usually has a leakage resistance between 100 gigaohm and infinity. A good vintage paper and oil capacitor, on the other hand, often only reaches a few tens of megaohm, and in some cases even less, with values that can be unstable over time. This is a significant risk when using a paper and oil capacitor as an interstage coupling capacitor. It can shift the grid bias of the downstream tube, with potentially disastrous consequences.

I am not saying this just to be a “prophet of doom” or to draw attention. On the Emissionlabs tube manufacturer’s website, there is a technical note warning that if you use paper and oil capacitors as interstage coupling capacitors, they will not provide any warranty. In other words, if you fry a pair of 300B tubes worth 1500€, it is your fault.

Here is the link to the page on the Emissionlabs website…

Screenshot…

Translated into Italian: The use of paper and oil capacitors as coupling capacitors towards the grid of the tube. These capacitors are designed exclusively for motor control and are extremely effective at self healing. However, their drawback is that they may exhibit small leakages at unexpected moments. These small leakages or noises may appear and disappear. Although they do not create problems in motor control, they can cause a shift in bias that may damage the tubes. The use of such capacitors is not recommended and voids the warranty… (Note: And this is being said by one of the most expensive and prestigious tube manufacturers on the market, not by “poor” Stefano Bianchini from SB-LAB who supposedly understands nothing about tubes).

In the photo below, you can see a concrete example of what can happen when a paper and oil capacitor is used as decoupling between two stages. The picture was sent to me by a customer who owned an artisan built device, constructed on the basis of myths and legends rather than real technical criteria:

Let us look at some measurements taken with the gigaohmmeter on three types of capacitors: a polypropylene capacitor, a good quality US military paper and oil capacitor, a lower quality Russian paper and oil capacitor, also NOS, and finally a WESTERN ELECTRIC part.

  • The polypropylene capacitor… It measures infinity
  • Now a US military grade paper and oil capacitor, or at least I believe it is paper and oil. Let us say that it could be an extremely high quality paper and oil capacitor, or something unknown…

  • Now we move on to a NOS paper and oil capacitor of Russian origin and military production, often found at markets and online bazaars and highly sought after by audiophiles. Here, however, some surprises are hiding…
  • Now the Western Electric capacitor, a part that I could easily sell on eBay for 250€ in a matter of minutes…

As you have seen, paper and oil capacitors have a drawback that, in some cases, can lead to damage. My mentor G. Mariani (known for the GRAAF OTL amplifiers, class of 1935) always advised me not to use them in the signal path for any reason. If you want to take advantage of paper and oil capacitors in a tube amplifier, I recommend using them for supply decoupling, following the schematic you see here.

After everything I have shown you, one last issue remains. If you want to use NOS paper and oil capacitors in your projects, it is absolutely essential to check their health, that is, the level of insulation they can provide, ideally at their rated voltage. This will allow you to determine whether the capacitors are compromised or still usable. A gigaohmmeter like mine is quite difficult to find nowadays and has a limitation: the test voltage is fixed at 500 volts, which may be too high for many paper and oil capacitors rated at 300, 250, or even 200 volts, making measurement impossible. Perhaps in this case we might get some help from PierAisa with one of his kits.

Final Recommendations

Paper and oil capacitors are generally designed to be hermetically sealed. However, if you find a capacitor that is damaged or oily on the outside, it is absolutely essential not to handle it any further. In that case, I recommend washing your hands with an effective soap, such as dishwashing liquid, to remove any trace of toxic oil. It is then better to seal the capacitor in a bag and dispose of it safely. This precaution is important because I have seen people mount paper and oil capacitors that already appeared oily on the outside, a clear sign of possible leakage. As long as they remain sealed and intact, they are harmless, but the oil inside is extremely toxic and can pose a serious health risk.

In the photos below, I show what I found while disassembling a device that had been bought on a classifieds site like Subito.it or similar, built by an improvised hobbyist and sold for 5-600€, advertised as the pinnacle of high fidelity. In reality, besides being poorly assembled and with disappointing measured and sonic performance, the owner gave it to me for scrapping in exchange for a small discount on another job. Inside, I found two visibly damaged paper and oil capacitors, completely oily and with a strong smell of leaking oil. They were so compromised that the measuring instrument went full scale already in precharge mode…

guest

This site uses Akismet to reduce spam. Learn how your comment data is processed.

0 Commenti
Inline Feedbacks
View all comments