Hi, it’s a bit silly, but I’m trying to make a noninverting amp from CA339E comparator according schematic below.
As expected the OUT voltage is composed from triangle wave at frequency of about 500ns (swithing speed of comparator).

Two questions:

Why the OUT sine is shifted?
Is there a better way to do this?

 

Because you don't understand what you are doing.

 

Ya, now I got that, since the output filter time constant was too short the comparator switches between on/off with different duty ratio so the error becomes obvious. Changing to 100n solved it.

 

It's still going to act squirrely with varying Loads and varying Frequencies,
and will probably, eventually, over-heat and fail.

Get a proper Op-Amp specified for the particular application at hand.

Most high-quality Op-Amps will usually do a better job of replacing a Comparitor,
rather than the other way around,
depending, of course, on the application.
.
.
.

 

Are you familiar with feedback loop stability? The Nyquist criteria and all that jazz?

Do you understand that a comparator, without the means to implement feedback loop compensation, cannot be reliably used as a stable linear amp?