RC filter selection

10x the 3dB frequency won't give you 90 degrees; it'll give you 84 degrees, and a ton of attenuation. You'll never get to 90 degrees with a single stage RC, and the further you go up in frequency the worse the attenuation will get.

Cascade an RC and you'll get 90 degrees at f = 1/(2*pi*R*C), with a signal strength that's only been attenuated 3:1. Put that into the + end of a voltage buffer or positive-gain op-amp and you'll be fine.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

Hi,

Did you mean Cascade two RC circuits together one after another? please explain !

Thanks

John

Vin ___ ___ Vout o-----|___|---o---|___|----o------o | | | | --- --- --- --- | | | | === === GND GND (created by AACircuit v1.28.6 beta 04/19/05

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

Or use an all-pass. Then it's easy to get 90° ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
           Only as good as the person behind the wheel.

All-pass:

Vin ___ Vout o-----|___|---o---o | | --- --- | | o -Vin

Get -VIN (180° with an inverting OpAmp)

And no loss of amplitude. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
           Only as good as the person behind the wheel.

Hi,

High pass or cascading schemes. The other problem is that how should I chose the value of R for both cascading and non cascading because the load is an opamp with very high input impedance.

Thanks

John

Kewl. If Vout needs buffering it'll use an extra op-amp, but still neato.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

I didn't invent that. Not sure where I picked it up. But I've been using that scheme...two of them staggered to maintain 90° differential over about an octave (for image reject mixers/LO's), ever since my first GPS chip... over 20 years ago. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
     Performance only as good as the person behind the wheel.

Another approach might be to use an RC and a CR filter -- providing -45 and

+45 degree phase shifts -- and then let the op-amp subtract one from the other, giving you 90 degrees total.

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Gee, and I thought my "use an RC and a CR and then subtract" was vaguely clever -- here you're doing it via superposition, effectively. Very slick.

Although you presumably have to worry about the source impedance of Vin being rather lower than the resistor's for this to work well...

---Joel

Hello,

Thanks for the suggestions. Would you please explain a little more that how can I implement all pass filter. I am unable to understand the diagram that you sent me? Plus if I choose to go with the low pass than I am still unable to undertsand that how will I pick the resistor value.

John

Naaaah! In the GPS, it's PECL followers. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
     Performance only as good as the person behind the wheel.

The rectangular box in his diagram there is the resistor.

You pick the resistor value based off of setting 1/(2*pi*R*C) to be the frequency you're interested in obtaining a 90 degree phase shift at and solving for R. Pick a "reasonable" value for C -- e.g., 1uF or so -- and then calculate R... if it's 100kohms ("reasonable" resistor values), change your capacitor's value to compensate.

People are likely ignoring you a bit here since this sounds suspiciously like a homework problem. If it is, and if you use Jim's method there (and please actually understand how it works), your instructors should be impressed.

(Granted, in the real world you might not have "-Vin" readily available, so you might want to come up with the "textbook" solution as well in case they're picky and there are unstated restrictions to the problem... such as, "your answer must match the one the professor is expecting." :-) )

Go download a copy of LTspice and verify your solution if you're a little unsure of yourself.

---Joel

John1987, my suggestion is that you stick with the two series low passes. I think you understand it. 10k ohms is a good starting resistor value for a lot of opamp circuits. Calculate what C you need and go on from there.

George H.

Google "all pass filter" and opamp if you want to go that way.

Hi,

Thanks for the replies. Honestly this is not a home work. I am not good at Analog design at all and learning the ropes. How does the bandpass filter work with Vin positive and Vin neagtive proposed by Jim.

John

Allpass, not bandpass.

-Vin => Vin _inverted_ by using an OpAmp with unity-gain-inverting feedback.

To understand how the R-C network makes an AllPass you'll need the capability to do Laplace math (at least in the "Laplace notation" form developed by Heaviside:

)

-or- the ability to do complex math (j-Omega-C, etc.) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
     Performance only as good as the person behind the wheel.

Do you understand how superposition works? If so, the intuitive way to view that circuit is that with the "-Vin" side set to ground, the output from "Vin" is your standard low-pass RC filter (...with 45 degree phase shift at

1/(2piRC), etc.). Then, with "Vin" set to ground, the output from the "-Vin" side is your standard high-pass CR filter. The overall output is the sum of these two.

Alternatively, you can just crank through the algebra... are you familiar with phasors (or Laplace transforms) and Kirchoff/Noton/Ohm's laws?

(I have a book aimed at electronic techs getting 2-year degrees, and their insistance on avoiding phasors makes a lot of the problem much more difficult; I felt they were doing their students a disservice, as I think you can teach people how to *use* phasors in just a couple of classes, which is of great utility even if they don't fully understand the bit of math behind it or the connection with Fourier/Laplace transforms.)

---Joel

If there's an inverting amp allowed, an integrator with a (simple darlington?) modest-gain amplifier would do the job. I'd dislike paying for a good op amp just to massage sinewaves at 200 kHz.