sonicap capacitor

For a 1uF, 10% rated cap, anything from 0.9uF to 1.1uF is within tolerance.
With film caps, the smaller values are cheaper to make, so they tend to be a little under the nominal value, rather than over.

Well I thought I'd test this hypothesis in a small way. Dug out all the 100nF film caps in my stock, which happened to all be "104 J100" (ie 5% tolerance) and measured with a tweezer multimeter of uncertain spec (but definitely repeatable).

The caps were in three shades of yellow, indicating 3 different manufacturers.

I had 3 of the dark yellow, mean = 111.4nF, s.d. 0.1nF
I had 4 of the lemon yellow, mean = 100.7nF, s.d. = 0.72nF,
And 12 of the medium yellow, mean = 98.37nF, s.d. = 4.76nF (min 91.5, max 108.6)

So clearly the first point to note is that many were simply out of tolerance completely.... And the worst offenders were _over_ value by 11% but very close in value (yes 3 is a small sample, but they were matched to 0.1nF...)

The medium yellow had the worse spread by far though, arguably more important.

However uncertainty in the measurement device might be misrepresenting things a little, but the overall spread of 91.5 to 111.5nF is clearly more like 10% than 5%.

Some of these are undoubtedly eBay purchases, if not all, generic stock for tinkering, likely PP. Not sure I have large sets of caps with good provenance, I tend not to over-order film caps for building pcbs as they are not prone to failure.