Note: A version of the material in Appendix I and II was part of a series of two articles I wrote for the Fall and Winter 2011 issues of 'The Antique Motorcycle,' the journal of the Antique Motorcycle Club of America.

Appendix I: Post-WWII Magneto Condensers

The Condenser

For the reasons I explain below, if you have a post-WWII British bike, either your magneto's condenser already has failed, or it soon will fail. When the condenser does fail, there are at least a dozen different categories of replacements (tantalum, metalized polyester, ceramic,…) made by dozens of manufacturers that have the necessary capacitance and will fit in the available space in the armature. Unfortunately, despite anything you might have been told before, essentially all of these will fail in operation, because most lack the ability to handle high current pulses. However, as I will describe in a subsequent post, I have conducted a series of tests that identify the type of replacement condenser that will provide many years of service.

Symptoms of a Bad Condenser

Often people who have a malfunctioning magneto say "I need to send it in for a rewind." In my experience, once other potential sources of problems have been eliminated (most commonly, a fouled plug, cracked high tension lead, short in the cutout circuit, or a worn or seized high tension pickup), in nearly all cases the problem is a bad condenser, not a faulty coil (unless it is a replacement rewound coil, in which case the coil could have developed an internal short). If your engine runs fine when cold, but misses and backfires heavily as it warms up, odds are high it is due to a bad condenser. Obvious sparking at the contact breaker points is the "smoking gun" of a faulty condenser.

Why all post-WWII Lucas and BTH Condensers Have Failed, or Are About to Fail

In 1915 B. H. Davies wrote "Riders need not worry about the action of the condenser, which never gives any trouble." In fact, this was largely true at the time he wrote it. Unfortunately for us, pre-WWII mica condensers gave way to paper condensers that were developed in no small part because of the disruption in the supply of mica from Zimbabwe (Rhodesia) and India caused by WWII. Also, while mica condensers don't suffer the specific degradation problem described below, they can fail from mechanical delamination or corrosion of the leads.

It turns out condensers (capacitors) of identical materials and construction as the ones Lucas and BTH used in their post-WWII magnetos also were of much interest to the telecommunications industry. Because of this, the greatest solid state physics laboratory of the time, Bell Telephone Laboratories, researched these "impregnated paper capacitors" in great detail. A 1946 book by a Bell Laboratories scientist, along with eleven research papers on paper capacitors published between 1942 and 1961, contain information that fully explains the cause of the problems post-WWII British magnetos are facing today (the book and papers are listed in the References section at the end of this post).

The reason these particular capacitors merited so much research was they filled an important niche. Each of the millions of phones Bell Telephone supplied to its customers contained one of these impregnated paper capacitors as part of the circuit that made the phone ring, so cost and longevity of individual components was a major concern. Paper capacitors, impregnated with wax, were much cheaper to produce than any alternative, including mica, and they worked reasonably well for as long as they worked. They weren't needed to last forever, but only long enough to span the gap until the next generation of telecommunications equipment was deployed. Because of this, much effort went into finding chemicals to add to the wax to ensure that essentially none of the capacitors would fail for at least a decade. While it was essential that no capacitor fail for at least 10 years, achieving longevity beyond 20 years for a significant fraction of them was not of much concern for manufacturers of these paper/wax capacitors, since that was beyond the planned service life of the equipment that made use of them.

A perfect paper/wax capacitor would consist of thin sheets of metal foil separated by thin sheets of paper soaked with some appropriate wax whose dielectric constant is as large as possible and whose electrical resistance is infinite. When it is fresh, chlorinated naphthalene -- trade names included Halowax, Seekay wax, and Nibren -- works quite well as that wax. As an aside, this substance is a PCB, which is now internationally banned because it is carcinogenic. Although the electrical resistance of this wax is not infinite at room temperature, and decreases rapidly with increasing temperature, it still remains high enough when the wax is fresh not to result in unacceptable capacitor performance. However, even with the best chemical stabilizers, the wax still degrades with time, although accelerated tests showed it would be good enough for the required decade of service.

The following photograph and magnified inset shows the paper and wax layers in a Lucas condenser.

[Linked Image]

There are ~125 layer pairs of area ~1"x1-1/2", with a separation between metal foils of ~0.001". Waxes have dielectic constants in the range 2.1-3.1. Using a value of 3 for an estimate, capacitance = dielectric constant x permittivity of vacuum x Area/separation x 125 pairs = 0.11 uF. The actual capacitance from Lucas literature is 0.15-0.18 uF, agreeing very well with this estimate based on my measurements of the internal structure of the condenser.

Even with stabilizers, research showed it also was essential to hermetically seal the capacitors because the wax is somewhat hydroscopic, and moisture accelerates the breakdown. When the wax breaks down it releases hydrochloric acid which then attacks the aluminum sheets of the capacitor, releasing aluminum chloride. Unfortunately, aluminum chloride accelerates the breakdown of the wax further, in turn releasing even more HCl. While breakdown of the wax happens no matter what, the process rapidly accelerates in the presence of moisture. The slight statistical variation in the permeability to moisture of the plastic seals is why Lucas and BTH magneto condensers fail over a range of ages. However, as the Bell Labs aging tests showed, even if you found a perfectly sealed new old stock Lucas or BTH condenser from the 1960s. it now would be approaching its maximum lifetime due to the chemical breakdown of the chlorinated naphthalene wax that happens even without moisture, and even if the condenser has never been used.

A few years ago I was given a truly new old stock Lucas condenser. The condenser was sealed in thick wax paper, in a cardboard box, in turn sealed in thick wax paper, and finally wrapped in paper on which was printed the part number and manufacture date of September 1956. All of the layers, including the outermost paper, were in fine condition. However, when I measured the electrical properties after extracting it, the capacitance was 0.601 uF (~4x larger than when new), and the dissipation factor was 15x larger than when new. This condenser would not have functioned if installed in a magneto, so paying a lot of money to buy a new old stock one on eBay is a very bad idea.

For a magneto, the relevant electrical consequence of the breakdown of the wax is an increase in the Equivalent Series Resistance (ESR) of the condenser. The condenser is connected in parallel with the contact breaker points specifically to provide a low AC resistance bypass for the current, i.e. to suppress arcing. While a perfect condenser would have ESR = 0 Ohms, as the ESR increases due to the growing electrical losses in the wax, the condenser's effectiveness as a bypass decreases, and the arcing increases. Typically, the first sign of problems is a magneto that functions acceptably when cold, but fails when warm. The reason for this is the electrical conductivity of the deteriorated wax changes exponentially with temperature. As a result, at this point in the life of the condenser, when cool the ESR is still low enough for the magneto to function, but when warm it becomes too high to suppress arcing. Whether the condenser is used or not, the wax will continue to deteriorate with time, and the ESR soon will be too high for it to suppress arcing even on the chilliest night. Confirming this behavior, I have measured over 50 Lucas paper/wax condensers, and their ESR values neatly follow a smooth curve that allows me to calculate how much longer any particular still-functional condenser will continue to suppress sparks.

The results of accelerated testing were known in 1946, so it is quite likely that Lucas and BTH were aware at the time these paper/wax condensers would begin failing in significant numbers starting in a decade or two. But, they also knew only a fraction of vehicles would remain in use after that many years anyway. And, when the condensers did fail, they could be replaced with no more cost and effort than, say, fitting a worn engine with a new set of rings. Also, although mica was again readily available, the cost to make a condenser with it was 20x higher than that to make one of paper/wax. In an industry driven by customers who made their purchases based on "value" (i.e. cheapness), and where warrantees expired after one year, their choice to use paper/wax was quite reasonable.

References

B.H. Davies, The Modern Motorcycle: How to Run, Ride, and Repair It (C. Arthur Pearson, London, 1915).

M. Brotherton, Capacitors: Their Use in Electronic Circuits (D. van Nostrand, New York, 1946).

D.A. McLean, L. Egerton, G.T. Kohman, and M. Brotherton, Paper Dielectrics Containing Chlorinated Impregnant: Deterioration in D.C. Fields. Industrial and Engineering Chemistry vol. 34, p. 101 (1942).

D.A. McLean and L. Egerton, Paper Capacitors Containing Chlorinated Impregnants: Stabilization by Anthraquinone. Industrial and Engineering Chemistry vol. 37, p. 73 (1945).

L.J. Berberich, C.V. Fields, and R.E. Marbury, Characteristics of Chlorinated Impregnants in Direct-Current Paper Capacitors. Proceedings of the I.R.E., p. 389 (June 1945).

L. Egerton and D.A. McLean, Paper Capacitors Containing Chlorinated Impregnants: Mechanism of Stabilization. Industrial and Engineering Chemistry vol. 38, p. 512 (1946).

D.A. McLean, Paper Capacitors Containing Chlorinated Impregnants: Benefits of Controlled Oxidation of the Paper. Industrial and Engineering Chemistry vol. 39, p. 1457 (1947).

L.J. Berberich and Raymond Friedman, Stabilization of Chlorinated Diphenyl in Paper Capacitors. Industrial and Engineering Chemistry vol. 40, p. 117 (1948).

J.R. Weeks, Metallized Paper Capacitors. Proceedings of the I.R.E., p. 1015 (September 1950).

H.A. Sauer, D.A. McLean, and L. Egerton, Stabilization of Dielectrics Operating under Direct Current Potential. Industrial and Engineering Chemistry vol. 44, p. 135 (1952).

D.A. McLean, H.A. Birdsall, and C.J. Calbick, Microstructure of Capacitor Paper. Industrial and Engineering Chemistry 45, 1509 (1953).

L. Borsody, New Impregnation for Paper Capacitors. IRE Transactions on Component Parts, 15 (March 1960).

Paul D. Garn, Stabilization of Capacitors. Industrial and Engineering Chemistry 53, 311 (1961).