"Slepian's Box"

That's a cool circuit, handy for all sorts of things.

It can be used to combine fast and slow amplifier paths into an output. Or make a good 50 ohm wideband termination when driving a capacitive load.

Never heard that name. Or any name.

The time-domain proof is to apply a 1V step voltage input and then look at the voltages across the resistors. One is going from 0 to 1V with tau L/R, and the other is going from 1v to 0 with tau RC. When the taus are equal, the sum of the two currents is constant.

I call that a Zobel network, (with an umlaut over the e) But I found the circuit before the name. I use it to "balance" coils, (with opamp-> fet/npn drive) To be honest I've never looked in much detail. For single supply I've used Al electro's in the cap branch.

George H.

Oh I always remember the formula as equal time constants. L/R = RC GH.

With input through the R's? Is it a "better" magic tee?

That's similar to making a coil look like a resistor. (Everyone loves resistors...)

George H.

Oops, I should have read to the bottom of your post, or I wouldn't have made the time constant comment.

Say this would work just as well to take apart a signal... and send the fast part over there... and DC this way. (I've got a control loop that would like that... I did it with one RC and instrument amps.)

George H.

That would give R^2 = L/C. The OP wrote it as R^2 = L*C. Is the OP's formula wrong or are they two different concepts?

I think Tim must have gotten it backwards. I know that if I find that the needed capacitor is too big, I make it smaller by increasing the resistance in the L branch. C=L/R^2

At one time in the long distant past I wrote out all the equations and solved it. You could do that? (or just do a dimensional analysis check)

George H.

I just know it as the standard second order diplexer circuit that you use folllowing a diode mixer or phase detector, to reduce back-reflections that cause IMD.

There are also higher order ones with the same property, I believe.

Cheers

Phil Hobbs

On the S.E.D/Schematics Page of my website...

...Jim Thompson

Oops.

How's that for an answer to your question?

L*C has dimensions second^2, which I should have twigged on when I wrote my original post.

And it's L/C = R^2. D'oh.