Eh, I'm not so sure if the usage characteristics of the machine has to do with leaky cans. Perhaps dried out or vented ones, but I would think the same chemical reaction that dissolves traces also dissolves the seals on the capacitor itself over time, regardless of use. Most of the machines we are repairing today haven't been in use for 20 years, yet still suffer. I'm not an expert in capacitor manufacturing, this is just personal observation.
Apple used plenty of tants over the years. There's *far* more damage done by electrolytics than tants shorting. It'd be hard pressed to find a single instance of an exploded tant on a vintage mainboard, where conversely thousands of boards suffer trace damage from leaky cans. And I seriously doubt anyone, even the most seasoned audiophile, could detect any difference in a blind A B between a tantalum and electrolytic on a vintage mainboard.
But I guess if that seriously concerns you, there are solid polymer electrolytic caps available, at a premium, that resolve those "issues".
Heat has quite a bit to do with the lifespan of the capacitors in question.
https://en.wikipedia.org/wiki/Aluminum_electrolytic_capacitor
The testing time and temperature depend on the tested series. That is the reason for the many different lifetime specifications in the data sheets of manufacturers, which are given in the form of a time/temperature indication, for example: 2000 h/85 °C, 2000 h/105 °C, 5000 h/105 °C, 2000 h/125 °C. This figures specifies the minimum lifetime of the capacitors of a series, when exposed at the maximum temperature with applied rated voltage.
Referring to the endurance test, this specification does not include the capacitors' being loaded with the rated ripple current value. But the additional internal heat of 3 to 10 K, depending on the series, which is generated by the ripple current is usually taken into account by the manufacturer due to safety margins when interpreting the results of its endurance tests. A test with an actual applied ripple current is affordable for any manufacturer.
A capacitor's lifetime for different operational conditions can be estimated using special formulas or graphs specified in the data sheets of serious manufacturers. They use different ways achieve the specification; some provide special formulas,[53][54] others specify their capacitor lifetime calculation with graphs that take into account the influence of applied voltage.[39][55][56] The basic principle for calculating the time under operational conditions is the so-called “10-degree-rule”.[57][58][59]
This rule is also well known as the Arrhenius rule. It characterizes the change of thermic reaction speed. For every 10 °C lower temperature, evaporation halves. That means for every 10 °C lower temperature the lifetime of capacitors doubles.
L x = L Spec ⋅ 2 T 0 − T A 10 {\displaystyle L_{x}=L_{\text{Spec}}\cdot 2^{\frac {T_{0}-T_{A}}{10}}} L_{x}=L_{{\text{Spec}}}\cdot 2^{{\frac {T_{0}-T_{A}}{10}}}Lx = life time to be estimated
LSpec = specified life time (useful life, load life, service life)
T0 = upper category temperature (°C)
TA = temperature (°C) of the case or ambient temperature near the capacitor
If a lifetime specification of an electrolytic capacitor is, for example, 2000 h/105 °C, the capacitor's lifetime at 45 °C can be "calculated" as 128,000 hours—roughly 15 years—by using the 10-degree-rule. Although the result of the longer lifetime at lower temperatures comes from a mathematical calculation, the result is always an estimation of the expected behavior of a group of similar components.
Capacitor seals are chemically resistant to the liquids used (circuit boards are not). I have seen blown tants, last one being on a Daystar 040 PDS board.
