How do you tell what type a capacitor is, (e.g. MKT, MKP, Mylar)? Most capacitors probably don't have a datasheet. Also, if the capacitor has no markings/code, for example:
is it possible to guess the type of the capacitor?
6
\$\begingroup\$
\$\endgroup\$
-
7\$\begingroup\$ Most capacitor manufacturers do produce datasheets for their capacitor families or styles (but not for individual capacitors). \$\endgroup\$ – Peter Bennett Jun 13 '16 at 22:30
2
\$\begingroup\$
\$\endgroup\$
Sparky256: I just did this:
Might be helpful. You can tell polyester from polypropylene, and maybe polycarbonate (rare these days).
-
\$\begingroup\$ Thanks for the info, especially since it was a DIY effort. I seldom dig into the 'old stuff' box for capacitors anymore as some of them are decades old. They are so cheap now I buy new capacitors, including smd types, as my DIY projects are becoming more smd based. The bags are labeled in detail in terms of composition, etc. \$\endgroup\$ – Sparky256 Jul 16 '16 at 20:15
9
\$\begingroup\$
\$\endgroup\$
There is NO solid answer to your question.
The capacitors shown look to be Mylar (reddish-brown), but I have also seen them with dark green coatings. I have some large reddish-brown 1uF 630vdc capacitors with the same body shape, but they are made with a polystyrene dielectric.
Depending on shape and size and manufactures specs, the wrapping or coating may change to accommodate assembly procedures and how the capacitor is sealed against moisture. I am used to ceramic disc capacitors having a light brown color, but that is not a standard set in stone. Ceramics can be small block-style and colored yellow, or dark brown for smd versions.
I have used large 3KVDC .15uF polypropylene capacitors made by GE that were wrapped in yellow, but without the datasheet and the fact that I ordered them as a pulse filter, I would have no way to know the materials used. At least these had part numbers on them, but most capacitors on a PCB do not have room for more than a capacitive code and voltage-if that.
- The bottom line is that if the coating is not obvious from constant usage of that type of capacitor, and there is no part number, your out of luck. No practical way to find out unless you cut one open and use chemicals to test what the dielectric is made of.
EDIT: As shown by Robert Endl's testing for dissipation factor with different dielectrics, if one has the need to, a DIY chart can be created that shows how the dissipation factor changes for a given dielectric type, all having the same capacitive and voltage value. This would be helpful for capacitors out of a junk bin that can be sorted with a good LCR meter.
Of course the chart must be created first with known dielectrics before sorting unknown dielectrics can be done.
-
\$\begingroup\$ A decent LCR meter might work. The dissipation factor at 1KHz vs 100KHz should be different for PET vs BOPP for example. Those are the two main materials. \$\endgroup\$ – Robert Endl Jul 13 '16 at 17:19
-
\$\begingroup\$ @RobertEndl. I agree that a good LCR meter would show differences in the dielectric, but one would need a chart of known values per capacitor value and dielectric type. If they exist, maybe just the manufactures have them, and do not publish certain spec's. \$\endgroup\$ – Sparky256 Jul 13 '16 at 21:02
-
\$\begingroup\$ You would have to calibrate the dissipation factor function vs capacitors with known dielectrics. Capacitance value should not be a big deal. It's not a great solution, but I think it's possible. \$\endgroup\$ – Robert Endl Jul 14 '16 at 21:29
-
\$\begingroup\$ @RobertEndl. I have a good Fluke LCR meter which could show the difference, but I would think of it as a DIY project. I have never seen templates that compare dielectrics vs. 1KHZ and 100KHZ readings. I know that for intense dv/dt pulses only polypropylene survives. Mylar will burn in the same test setup. \$\endgroup\$ – Sparky256 Jul 14 '16 at 21:36
-
\$\begingroup\$ Just the dissiparion factor at a given frquency might work. Polyester should be much higher that polypropylene for example Telling polystyrene from polypropylene might not work however. \$\endgroup\$ – Robert Endl Jul 15 '16 at 22:36
