passive phase shifter design?

[Ancient_Hacker]

Got an intersting problem:

inputs: sine wave, 50 to 500 Hz, 10 volts p-p, 50 ohms, and its 180
degree opposite.

Outputs: two sine waves, 90 degrees apart, within 2% in phase, 20% in
amplitude.
Working into a 100K ohm load, at least 2v p-p.

Catch: has to be done with just r's and c's, no active components.

I see lots of phase-shifters that use op-amps, or work at one
frequency, or work from 300-3000 Hz, but nothing in the range I need.

Any hints appreciated.

 

[Tom Bruhns]

Got an intersting problem:

inputs: sine wave, 50 to 500 Hz, 10 volts p-p, 50 ohms, and its 180
degree opposite.

Outputs: two sine waves, 90 degrees apart, within 2% in phase, 20% in
amplitude.
Working into a 100K ohm load, at least 2v p-p.

Catch: has to be done with just r's and c's, no active components.

I see lots of phase-shifters that use op-amps, or work at one
frequency, or work from 300-3000 Hz, but nothing in the range I need.

Any hints appreciated.

If you have one that you're satisfied with from 300-3000Hz, then just
scale the capacitors up by 300/50 = 6. Or scale the resistors down by
that factor, or do a combination of the two.

Sorry, that was more than a hint.

Another non-hint: look for designs of phasing-type SSB
encoders/exciters/decoders/detectors. They must somehow accomplish the
same sort of thing. Note that the phase of the output will NOT be
constant relative to the input, just 90 degrees apart between the two.

Cheers,
Tom

 

[Ancient_Hacker]

Thanks Tom, that is the hint I needed. The old ARRL handbooks show
audio phase shift networks for the "communications" audio range. I
just didnt know whether to scale the resistors, the capacitors, the
impedances, or some combination of those.

Also found some "polyphase" networks that are supposedly less tolerance
sensitive.

Thanks,

grg

 

[Jim Thompson]

Got an intersting problem:

inputs: sine wave, 50 to 500 Hz, 10 volts p-p, 50 ohms, and its 180
degree opposite.

Outputs: two sine waves, 90 degrees apart, within 2% in phase, 20% in
amplitude.
Working into a 100K ohm load, at least 2v p-p.

Catch: has to be done with just r's and c's, no active components.

I see lots of phase-shifters that use op-amps, or work at one
frequency, or work from 300-3000 Hz, but nothing in the range I need.

Any hints appreciated.

(1) Broadband phase shifters are non-trivial. Surf old Bell Labs
journals, Darlington... yep, the same guy of compound transistor fame.

(2) If you can adjust at a single frequency...

R/C between VIN (0°) and -VIN (180°) can be set to almost any phase
angle at the tap between the R and C... amplitude is constant!

...Jim Thompson

 

[Simone Merrett]

Got an intersting problem:

inputs: sine wave, 50 to 500 Hz, 10 volts p-p, 50 ohms, and its 180
degree opposite.

Outputs: two sine waves, 90 degrees apart, within 2% in phase, 20% in
amplitude.
Working into a 100K ohm load, at least 2v p-p.

Catch: has to be done with just r's and c's, no active components.

I see lots of phase-shifters that use op-amps, or work at one
frequency, or work from 300-3000 Hz, but nothing in the range I need.

Any hints appreciated.

Can inductance be used?
Why Not Active Components?
What about non-linear Components? (EG Diodes, transistors?)

Assuming the limitation is because it cant use separate power source, It
would be quite simple to suck a bit of power from the sine input to drive
some active components. (EG A digital quadrature divider... which is quite
simple with a couple of 74series ICs.)
How "sine" does it have to be? (is a sq wave ok?).

It may be possible with an analog feedback system of L, C & R &
transistors, so that feedback biases so as to null a multiplied sample of
the outputs. (multiplying 2 sines 90Deg apart produces A DC of 0V, and
deviations in phase from 90Deg produce DC..)

David Merrett