Not too long ago, I came across a few very unusual capacitors. That got me thinking about some other weird components in my stash.
These small, weird caps were gifted to me as part of a junk box no longer needed. Looking at the unusual packages, I wasn't even sure they were caps. My ohmeter said open. My capacitance meter said about 100 nF, but the reading was not stable. I eventually realized that the reading was highly temperature-sensitive: somewhere around -6,000 ppm/K. Was this an unusual sensor?
Help from an email list turned up a promising Siemens databook page, but even that didn't fully solve the mystery. One of the tables (in German, by the way) lists several N- and P-type ceramics. You may not know, but ceramic caps are available with controlled temperature coefficients, which are commonly used to compensate for another unstable component, like a coil. Tempcos from P100 (+100 ppm/K) to N5000 (-5,000 ppm/K) are typically seen. My cap is an S type. What's that? The table shows S caps with tempcos from 700 to 5,000. Is the function nonlinear? S-shaped, perhaps? The mystery is semi-solved. Do you know more?
Next up is another cap. It's not quite as weird as the first, and it appears to be a 100µF 10V tantalum. The only really weird thing is its shape. It's rectangular — or, should I say, prismatic. Have you seen anything like it before?
The third and last weird cap in my stash was designed to provide the ultimate in low-impedance filtering for switching power supplies. It's just an electrolytic — a four-terminal electrolytic, that is.
The idea is that two leads at one end are the input, and the other two are the output. I assume that each wire runs continuously through the length of the package, contacting the physical capacitor along its entire length. This eliminates most of the resistance and inductance that always gets in between the PC board and the actual capacitor.
To read the rest of this article, visit EBN sister site EDN.