Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

73, Fred, K4DII

Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

--
All relevant people are pertinent.
All rude people are impertinent.
Therefore, no rude people are relevant.
-- Solomon W. Golomb

Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

--
All relevant people are pertinent.
All rude people are impertinent.
Therefore, no rude people are relevant.
-- Solomon W. Golomb

"Frank Mikkelsen" wrote in message ...
I whant to double a frq. between 200 and 265 MHz. I had tryed to used the
Mini-circuits RK2 doubler and som BP filters, but because of the band pass
200-265MHz it is diffeculd for me to suppress the input frq. and the 3*inp.
to min 50-60dB.
Is there any other ideas to realise this.

I think the doubler may be the easiest way to do it. Because your
output frequency range is so limited, it should not be difficult to
build a bandpass filter for that range, perhaps even with "zeros" on
the input and on 3*input. Even without the zeros, a simple
5-resonator 0.1dB ripple Chebychev design from RFSim99 shows 30dB
atten at 300MHz, and that's the worst-case 1*input, 3*input atten.
Admittedly, you won't achieve quite that much in a practical
implementation, but should be close. (Quick playing with a zero at
300MHz makes me think a 3-resonator filter with that added zero might
be enough.) Add that to the rejection of 3rd you should already be
seeing (30dB min according to MiniCkts) and that should do the trick
for you. Also, you can "trim" a doubler to have high rejection of the
third, and low feedthrough of the fundamental, if you want.

You may also find some interesting ideas on the Wenzel web sites:
specifically at http://www.wenzel.com/pdffiles/diodedbl.pdf, but you
may find other things under the RF Circuits section of
http://www.techlib.com/electronics/index.html.

Cheers,
Tom

"Frank Mikkelsen" wrote in message ...
I whant to double a frq. between 200 and 265 MHz. I had tryed to used the
Mini-circuits RK2 doubler and som BP filters, but because of the band pass
200-265MHz it is diffeculd for me to suppress the input frq. and the 3*inp.
to min 50-60dB.
Is there any other ideas to realise this.

I think the doubler may be the easiest way to do it. Because your
output frequency range is so limited, it should not be difficult to
build a bandpass filter for that range, perhaps even with "zeros" on
the input and on 3*input. Even without the zeros, a simple
5-resonator 0.1dB ripple Chebychev design from RFSim99 shows 30dB
atten at 300MHz, and that's the worst-case 1*input, 3*input atten.
Admittedly, you won't achieve quite that much in a practical
implementation, but should be close. (Quick playing with a zero at
300MHz makes me think a 3-resonator filter with that added zero might
be enough.) Add that to the rejection of 3rd you should already be
seeing (30dB min according to MiniCkts) and that should do the trick
for you. Also, you can "trim" a doubler to have high rejection of the
third, and low feedthrough of the fundamental, if you want.

You may also find some interesting ideas on the Wenzel web sites:
specifically at http://www.wenzel.com/pdffiles/diodedbl.pdf, but you
may find other things under the RF Circuits section of
http://www.techlib.com/electronics/index.html.

Cheers,
Tom

"Clifton T. Sharp Jr." wrote in message ...
Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

Well, actually the whole output wave is at 2* the input, which is I
believe what the OP wanted: a freq doubler. Note that rectifying a
sinewave is the same as multiplying by a square wave of the same
frequency which is +1 when the sine is positive and -1 when it's
negative. You can use trig identities and the fact that the square
wave is its fundamental and all odd harmonics to convince yourself
that the "ideal" full wave rectifier puts out DC, 2*fin, 4*fin, 6*fin,
.... -- just the even harmonics and no odds. It's only because the
rectification is imperfect that the fundamental or odd harmonics get
through. The pointy lower peaks must represent higher order even
harmonics of the input frequency.

Cheers,
Tom

"Clifton T. Sharp Jr." wrote in message ...
Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

Well, actually the whole output wave is at 2* the input, which is I
believe what the OP wanted: a freq doubler. Note that rectifying a
sinewave is the same as multiplying by a square wave of the same
frequency which is +1 when the sine is positive and -1 when it's
negative. You can use trig identities and the fact that the square
wave is its fundamental and all odd harmonics to convince yourself
that the "ideal" full wave rectifier puts out DC, 2*fin, 4*fin, 6*fin,
.... -- just the even harmonics and no odds. It's only because the
rectification is imperfect that the fundamental or odd harmonics get
through. The pointy lower peaks must represent higher order even
harmonics of the input frequency.

Cheers,
Tom

Tom Bruhns wrote:
"Clifton T. Sharp Jr." wrote in message ...
Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

Well, actually the whole output wave is at 2* the input, which is I
believe what the OP wanted: a freq doubler. Note that rectifying a
sinewave is the same as multiplying by a square wave of the same
frequency which is +1 when the sine is positive and -1 when it's
negative. You can use trig identities and the fact that the square
wave is its fundamental and all odd harmonics to convince yourself
that the "ideal" full wave rectifier puts out DC, 2*fin, 4*fin, 6*fin,
... -- just the even harmonics and no odds. It's only because the
rectification is imperfect that the fundamental or odd harmonics get
through. The pointy lower peaks must represent higher order even
harmonics of the input frequency.

http://www.rfcafe.com/references/electrical/periodic_series.htm has
a nice exposition of Fourier series for various waveforms, including
half- and full-wave rectified sine waves. The Rubber Bible math table
book used to have the waveforms and Fourier series for them, too.

I first got interested in them when I was about 11, in 1957. Somewhat
later I took the math class where we derived them. It was sort of
interesting to see my childhood friends constructed on a blackboard.

--
Mike Andrews

Tired old sysadmin since 1964

Tom Bruhns wrote:
"Clifton T. Sharp Jr." wrote in message ...
Fred McKenzie wrote:
Full-wave unfiltered rectification, followed by bandpass filter?

Clifto-

Doesn't full wave produce a symmetrical waveform that minimizes even harmonics?

It's not exactly symmetrical. Rectifying a sine wave produces what looks
like a sine wave with pointy lower peaks, at twice the input frequency.
A little decent filtering at 2F, and the pointy lower peaks go away.

Well, actually the whole output wave is at 2* the input, which is I
believe what the OP wanted: a freq doubler. Note that rectifying a
sinewave is the same as multiplying by a square wave of the same
frequency which is +1 when the sine is positive and -1 when it's
negative. You can use trig identities and the fact that the square
wave is its fundamental and all odd harmonics to convince yourself
that the "ideal" full wave rectifier puts out DC, 2*fin, 4*fin, 6*fin,
... -- just the even harmonics and no odds. It's only because the
rectification is imperfect that the fundamental or odd harmonics get
through. The pointy lower peaks must represent higher order even
harmonics of the input frequency.

http://www.rfcafe.com/references/electrical/periodic_series.htm has
a nice exposition of Fourier series for various waveforms, including
half- and full-wave rectified sine waves. The Rubber Bible math table
book used to have the waveforms and Fourier series for them, too.

I first got interested in them when I was about 11, in 1957. Somewhat
later I took the math class where we derived them. It was sort of
interesting to see my childhood friends constructed on a blackboard.

--
Mike Andrews

Tired old sysadmin since 1964