Vintage Tubular Capacitors: Paper or Polymer?

When repairing vintage electronics, especially vacuum-tube gear of the 1960s and earlier, tubular capacitors are often encountered. These are simple components that consist of two metal electrodes rolled into a cylinder with a dielectric material between them. This dielectric is either paper or plastic film, or sometimes both wound together as in the case of Difilm capacitors. This capacitor roll is then encapsulated, either in hard plastic, cardboard sealed with wax, or metal cans sealed with rubber or glass.

In restoring an electronic device, it is valuable to know quickly and reliably what kind of dielectric a capacitor uses, in particular whether it is paper or not. Paper capacitors are notorious for developing electrical leakage and shorting out, and thus are often replaced preventatively, unless the type is known to be especially stable. Conversely, there is no reason to replace plastic film capacitors as a preventative measure, because all polymer dielectrics age very well in comparison to paper—only in the rare event of failure should they be replaced. I also tend to leave Difilm capacitors in place, since if the paper degrades, the permissible working voltage may decrease (and the capacitance may change), but at least the polymer prevents leakage from developing prior to a breach of the dielectric, which I've never seen happen so far.

There is a lot of misinformation about this topic online. I have read in multiple places that Sprague Bumblebees without spouts are mylar or difilm. They aren't. Every one that I've dissected was paper. I've even read that most molded-plastic capacitors in vintage tube electronics are plastic film, and that the polymer breaks down just like paper. This is also completely false; most of them are paper, with the transition to plastic film being made primarily in the early 60s, and those that are polymer do not develop leakage the way paper does.

Tired of misinformation, and more commonly a plain lack of information, I decided to make a list of my own. This list comprises only info I have either personally verified, or taken from official datasheets. By "personally verified", I mean that I have broken open one or more individual capacitors and looked inside. Since I keep most of the parts I replace, I had a fair number of capacitor types to check.

Also included are a few Soviet-made paper types, because they are quite common on eBay and similar sites. These were produced in great quantities well into the 1980s, long after paper caps were abandoned by the West. They frequently appear in Soviet-made tube and solid-state electronics. So, without further ado, here is the list:

Note that my list generalizes; I assume that all of a particular type of capacitor use the same dielectric, based on only one or a few examples. Also note that I do not distinguish between wax paper and oil-soaked paper, because both are generally prone to developing leakage as the paper decomposes and absorbs moisture. The only significant exceptions are the hermetically-sealed types, because the oil cannot evaporate nor react with air, and the paper cannot absorb moisture or react with air either. Some non-hermetic types also tend to age better than others, undoubtedly due to differences in paper type, impregnation, and "sealedness". For example, Sprague Bumblebees with spouts can be especially bad (probably due to the oil being filled via the spout only from one side), whereas I've had very good experience with Pyramid IMP molded paper types.

I should state plainly that I am not encouraging replacing all paper capacitors — in many cases, especially with relatively late molded types, they can be left in place without problems. Rather, my main intention is to save myself and other technicians valuable time, effort, and resources by pointing out which types are plastic film, and which types should thus be excluded from recapping.

If you have capacitors not listed that you are curious about, here are some suggestions. If you don't wish to destroy the capacitor, a nearly failproof test can be done using Mr. Carlson's Lab low-voltage leakage tester. The design is only available on his Patreon page, and you have to build it, but it performs magnificently. In particular, the extremely sensitive "forecasting / mica" setting will almost always detect leakage on a paper capacitor, even if it tests fine on the "paper" setting. Yet, it will always show no leakage on a polymer (or Difilm) capacitor, even those from the 1950s and 60s.

While the Carlson tester gives you a good idea of whether a capacitor is worth worrying about, breaking it open is the only way to be certain of what's inside. I find that squeezing between large pliers is sufficient to break away the shell of most molded plastic types, and metal cans can be pierced by flush cutters. Wear eye protection and disposable gloves, and work in a well-ventilated area, since you may encounter nasty oils, possibly including those containing PCBs. Once the shell is breached, the rolled layers can be cut with a blade and picked apart using fine tweezers.

From picking at the layers, you can easily discern the material. Paper is always translucent to opaque, and white to brown in color. It tears easily, giving a ragged fibrous edge. This is compared to polymer, which is transparent to translucent, and will sometimes stretch before tearing. When it tears, it does so in a clean line, or with some scalloped irregularity (if stretched to breaking), devoid of fibres. Most often, the dielectric will be easy to separate from the aluminum foil electrodes. Sometimes, the dielectric is "metallized", meaning it is coated in aluminum, comprising the electrode instead of a separate foil layer. It can be slightly harder to identify the material as a result of being unable to distinguish its color and opacity, but the same tests can be applied: stretching and tearing will reveal whether it is paper or film. Also, keep in mind that metallized polymer types are more common than metallized paper.

Below shows a metal-can paper capacitor cut open with flush cutters starting at the rubber seal. Note that the paper is brown, translucent, fibrous, and easily torn. This capacitor showed no leakage even on the "mica" setting of a "Mr. Carlson" tester, leading me to think it could be plastic. No, it was just an especially good paper-in-oil. The paper was still quite wet; the rubber seals really did their job on this capacitor.

Below shows the inside of a 1950s polymer capacitor: a Good-All 600UE. This tested perfectly in both capacitance and leakage. Breaking it open, the film is very obviously plastic, since it is translucent blue, and stretches slightly before tearing cleanly and without fibres.

Lastly, here is a dielectric sample from a classic yellow Astron capacitor, showing that it is Difilm. You can see the transparent plastic base layer, on top of which is a partially torn translucent paper layer. The plastic stretches and has no fibres, while the paper tears easily, giving a fibrous edge.

If you'd like to contribute to this list, I certainly welcome it, though there is a certain standard of proof necessary: either provide a scanned datasheet (or other "official" reference), or pictures of the capacitor both in unbroken and broken states. A few types that I am especially curious about are as follows:

• Good-All types 623 and CMI-1. My suspicion is that both are plastic film, but I am not sure.
• Mallory "blue molded" caps, as used in Fender amplifiers and some Conn and Thomas organs, among others (they tend to hold up very well, and are probably plastic-film).