”Sweet” 3-pole filter analysis help

[lion]

In an old preamp I build many years ago (to work with a digital delay), I have a passive filter in the dry path. It's simple but odd – and I can't remember how it, nor the exact component values, came about.



The thing is I really like what the filter does, but I'm puzzled by what it seems to do when I look at the RC values!

What I'm hearing switching the filter in and out, and with different settings of P1, is of course not a standard treble cut type effect by any means. It's just takes off a tiny bit of shrillness from a Strat bridge pickup (probably why I put it in initially). However the main percieveable effect is that it seems to soften the pick attack on the higher notes, almost like a bit of compression. In fact it seems to work more like a nice "comp/limiter" on high notes, than as a tone control. Which is why I really like it.

What puzzles me is, if it's at all possible that the HF roll off point with the filter as shown is down in the guitar range.

I can calculate single pole RC filters, but I've given up on multiple pole RC's (especially after PRR at one point said it's idiot math, and that he can NOT do it).

Or am I missing something completely?

Erik

NB: I know a number of guys here are very handy with SPICE/TINA/etc (which I'm sadly not). A simulation of the filter response would be interesting and helpfull I'm sure 

[PRR]

It is not a 3-pole filter. 1-pole at most.

You have ~~200K feeding it, 440K loading it, so what can those 7K even 40K resistors do? They are all "small" compared to all other impedances. You can replace them with shorts. In sim, same graphs but 1 dB higher (as expected taking 50K out of 500K total load).

So short all the little resistors, use one 1,120pFd cap, same for all practical purpose. 1,120pFd against ~~200K is around 750Hz. Upper fundamentals. Works mostly on string overtones.

Pot at zero, 1-pole with -3dB at 1.2KHz. Pot turned more-R, 2dB-3dB shelf above 1KHz. Subtle. Like you said: "a tiny bit of shrill".

[lion]

Thanks Paul - I get it. A bad/misunderstood implementation of a (not even close) 3-pole filter on my part.

Those multiple RC filter keeps me intrigued. I seem to remember 3-pole examples with increasing resistor values (10k/100k/1M) and proportional decreasing C's. More research and reading up needed I guess.

Erik

[ElectricDruid]

Those multiple RC filter keeps me intrigued. I seem to remember 3-pole examples with increasing resistor values (10k/100k/1M) and proportional decreasing C's. More research and reading up needed I guess.

Yeah, that works, mostly. Purists will argue, but it works. The "much larger" resistor for the next stage limits the loading of one stage on another, so you get something closer to buffered stages.

Tom

[lion]

Thanks Tom. In the end higher order filters are probably easier done with Salen-key, Butterworth etc active circuits, but I mostly find that for every active stage added there are "side effects" one way or the other. Could be a psychokological thing though 

Erik