+- 90 phase shifter

Unless you do it digitally that will only work over a limited frequency range. Check our Hilbert transform circuitry. Usually consist of lots of precision R's and C's. There are also tricks how to do those with switched capacitor filters. Mostly used in ultrasound, Radar and somewhat ancient ham radio gear.

--
Regards, Joerg 

http://www.analogconsultants.com/

Sloppy ASCII...

+Ein o | R | +----o OUTPUT | C | o -Ein

Is reasonably accurate for about an octave. And is constant amplitude.

(Generate -Ein with an inverting stage, such as an OpAmp.) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

If the input is just plain sine wave, then phase lock loop could provide very accurate 90 degree shift over wide range. So would integrator or differentiator circuit, however AGC might be required after.

Vladimir Vassilevsky DSP and Mixed Signal Designs

¨ Audio ?

Microwave ?

Coiled spring and ceramic mike ?>:-} ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | I love to cook with wine. Sometimes I even put it in the food.

The right length of coax? For just one frequency.

Sure, one be long piece of wire, or what ever the freq you are working with. calculate the wave length in free space and then apply the velocity factor to shorten that figure up a bit. Of course it depends on what you're using as a conductor, you''ll also need to worry about matching the line to avoid reflection distortion.

THen again, you could always use a delay line

Jamie

No, but yes, kinda.

Any circuit, even as simple as a single conductor distorts its input to some degree, adding, if nothing else, noise. As your circuit increases in complexity, it will have distortions in amplitude, and delay that vary based on the incoming signal.

Engineering is all about defining the parameters of that distortion that are acceptable to your application, described in one of several domains, and developing a circuit that meets those parametrs.

The circuit I think you're looking for is often called an "all passs" filter in the analog domain, or a Hilbert Transform.

For a signal of a single frequency, it is nothing more than a fixed delay. If your signal can take one of many frequencies, you need a delay that is variable based on frequency, such that the delay is always the time that represents 90-degrees at each frequency.

The implementation I'm thinking of has one input and two outputs, that are each delayed from the input, but one is delayed 90 degrees more than the other over some range of frequencies.

Putting it in the concrete -- if I want to create a single-sideband signal to carry Voice Modulation, I can use a phase splitter that gives me 0 and 90-degree shifts over a limited band from roughly 300 to 3000 hertz.

--
NK1G - Lawrence 
echo 'lawrenabae@abaluon.abaom' | sed s/aba/c/g

"greysky"

** Completely ambiguous question.

Why don't people just SAY what it is they are trying to do.

... Phil

Well, if it was just one frequency all you need is one inductor and one capacitor, not a whole PLL :-)

--
Regards, Joerg 

http://www.analogconsultants.com/

0

M.J. Gingell ?Single Sideband Modulation using Sequence Asymmetric Polyp hase Networks? Electrical Communication 48 21-25 (1973). See also

describes an approach that provides a close approximation to a 90 degree ph ase shift over an appreciable range of frequencies. It's in the second edit ion of "The Art of Electronics" by Horowitz and Hill (ISBN 0-521-37095-7) i n section 5.17 as a "phase sequence filter".

--
Bill Sloman, Sydney

Hey, could you stick a divider before the mixer, lock to 1/2 the freq. and get the full 180? (I guess you'd need a divider after the VCO too.)

George H. So would integrator or

Yeah, a long piece of wire with a tap... how can you make the velocity really low? For microwave frequecies, we had this trombone like thing at the FEL. (not cheap!!!)

George H.

Are you generating this sine wave? If so, just generate it in quadrature.

IIRC a biquad implementation of a bandpass filter has a 90 phase difference between the op amps. But that is a narrow band solution.

I recall a project where I needed a wide band 90 degree phase shift and ended up using a paper out of the IRE journal. (RE as in Radio Engineers). It was a SSB phase splitter filter.

The solution has to be all analog. It is simple to do any kind of phase shifts by dividers and logic; but this is against the rules of contest.

Vladimir Vassilevsky DSP and Mixed Signal Designs

Act ignorant and use a L-C network..

Too simple for management..

On a sunny day (Tue, 18 Jun 2013 13:17:47 -0700) it happened "greysky" wrote in :

Well, if you only need the +90 and the -90 degrees, THEN just use a transformer to make 180 degrees.

|>:-}

I use a differential version for all my image-rejection mixers... as in, for instance, the Garmin GPS chip. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.