Integrator Filter Question...

[Earthscum]

I was curious... with an integrator, the R and C set a corner frequency, right? So basically, it amplifies everything below the Fc at the maximum gain the op amp can go, and drops gain on everything above the Fc, correct?

So, say I have a 10k and a .022u, which should give me a Fc= ~725Hz. Assuming a parallel resistance to the cap it would be considered infinite, what happens if we lower that resistance to 10M (1000x gain)? 1M (100x gain)? etc... what happens to the Fc? Is there an equation, or addition to the typical 1/(2*Pi*RC)?

[PRR]

> with an integrator

A perfect integrator has infinite gain at zero frequency. No "corner".

You can't build a perfect integrator.

There are dozens of ways to build imperfect (but good-enuff) integrators.

So get specific. WHAT integrator plan are you thinking of?

If the classic inverting opamp, then what you call "corner" is no corner, just the frequency where output is equal to input. If you can snag a 20Hz-20KHz sine oscillator, a few minutes at the breadboard will clarify that..... but you need to short the integration cap every few seconds because the near-infinite gain at zero frequency means the least DC error accumulates to infinity and drives the opamp past a rail.

> Assuming a parallel resistance to the cap it would be considered infinite

Not understanding this passage. Draw/cite a picture.

[Earthscum]

If I take a resistor and dropped it parallel with the cap, what would it do to the frequency? If there's no resistor, we can assume there's infinite resistance.

Basically why I'm asking, is that I've been using a 4049 inverter, putting a pair of 27k's in front, and taking feedback through a .022u cap to the center. The stage before that is usually just a 10x inverter gain stage. They work great together (gain/buffer stage biases the second stage). I use this setup as a glitch-less driver for CMOS, a bass dubber, etc.  Basically, it is an integrator preceded by a gain stage. I can take just about any signal and it rounds it out smooth, and as close to the rails as possible. When I try it with guit-fiddle, I lose signal about halfway up the scale (when I'm driving, say a schmitt trigger). a pair of 22k's and a .01u works great for guitar.

Basically, I figured out I'm using an integrator to do alot of fun stuff, and originally set up just an LP t-filter until I figured out I can get what seems to be a brick wall effect by feeding the cap forward instead of to ground. That's why I'm wondering about what role a feedback resistor (parallel to the cap) would play on how this thing reacts. If there's an equation for it, that's what I'm after. I'm tired if fiddling about with filters on the breadboard when all I have to do is crunch a number or 3 and figure things out properly, so I've been trying to do the math instead of just popping in components and assuming that it's working right. Ya know what I mean? Playing is fun, but doesn't get me any farther ahead, and leaves me with giant piles of caps and resistors to put away, lol.

[PRR]

You know (or can learn) the basic formula for an inverting opamp.

You know (or can learn) the basic formula for reactance of a cap at a frequency.

Put the two together, you can rough-out the frequency response in an instant.

The basic integrator is infinity at DC, zero at infinite frequency, and about unity when R = Zc.

> a resistor and dropped it parallel with the cap

Then low-frequency gain is less than infinite.

But also note that your 4049 does NOT have infinite gain, and will fail to satisfy the "perfect opamp" gain formula when predicted gain is near or above the 4049's actual gain.

So down to a point, a cap in parallel with the cap may not change the low-frequency gain much. It will however help the 4049 find a DC operating point. (A perfect integrator with even non-infinite DC gain tends to drift until it hits a rail.)

[Earthscum]

So down to a point, a (resistor?) in parallel with the cap may not change the low-frequency gain much. It will however help the 4049 find a DC operating point. (A perfect integrator with even non-infinite DC gain tends to drift until it hits a rail.)

I apologize, I'm really bad about getting things from my head to words. I work great if others understand analogous description.

And, sadly I get it now... I was just over-thinking things. About time to kick through the filter topo and design sheets and see how much more I understand since the last time I sifted through them all. Thanks for replying.