Simulating an RL circuit

John, Are you just discovering that trick?

In general: R*C = L/R

I was using that to make inductive drivers for modems ~1977.

But it dates MANY years back to GenRad bridge designs.

...Jim Thompson

A filter with a perfectly constant input impedance?

I think it works for any LC pair with that ratio of values and 50 ohm resistors.

Doesn't have to be 50 Ohms.

See my other post... as long as: R*C == L/R

...Jim Thompson

I think I'd seen it a long time ago, and forgot. I started thinking about the step-response currents in each leg and slapped meself upside the head. Duh.

We just rediscovered it in the context of split-path amplifiers. As a practical matter, to make it really precise, one would probably have to use a little Coilcraft slug-tuned inductor as part of L, to tweak things up. There will be capacitive parasitics in L, too, a more serious issue if this is to be truly wideband, like DC to 1 GHz for starters.

John

Yep. Although the GenRad bridges used precision wire-wound resistors paralleled with carbon resistor plus a cap.

I think it's actually fairly good even at frequency extremes and slight mismatches.

...Jim Thompson

That's a two-pole RC diplexer, viewed from the source's perspective. Those are commonly used in receivers to ensure that the mixer sees a constant 50-ohm load, which is very important in maintaining mixer linearity.

Cheers,

Phil Hobbs

So now we can put an amplifier in each leg. It gets trickier if the AC path doesn't have a simple single-pole rolloff, but we can probably work around that.

John