Low frequency signal generator

[Rich Grise]

On Thu, 28 Jul 2005 11:31:29 +0100, Dirk Bruere at Neopax wrote:

Robert Baer wrote:
Dirk Bruere at Neopax wrote:

Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff. One solution is
to create a stereo track with one channel at f1 and the other at f2
where f1>f2, feed these into a summing op-amp and follow by some simple
filtering to get f1-f2.

That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.

Summing them won't work?

No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

Harmonics are also generated.

I wonder if that's really the case, if your analog multiplier
is "linear". It should be simple for someone who paid attention
in that class to answer this one with arithmetic.

The f1+f2 can be filtered out. It should be easy if f1,f2 >> f1-f2

True.

Cheers!
Rich

 

[Ben Bradley]

On Thu, 28 Jul 2005 08:22:54 -0400, Spehro Pefhany
<speffSNIP@interlogDOTyou.knowwhat> wrote:

On Thu, 28 Jul 2005 06:34:26 GMT, the renowned Robert Baer
robertbaer@earthlink.net> wrote:


In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.
Harmonics are also generated.

Some audio editing software allows you to generate a modulated signal.

You can easily enough (with appropriate software) generate a very
low frequency in a .wav file, but the earlier reason for doing it wbut
virtually all playback devices roll off below 20 Hz.

Here's a webpage on modifying a soundcard's input to respond down
to DC, perhaps its output could be similarly modified:

http://www.qsl.net/om3cph/sb/d cwithsb.htm

But I think that's a dodgy kludge either way.

Here's a 'real' interface with two analog outputs:

http://sine.ni.com/nips/cds/view/p/lang/en/nid/14604

Best regards,
Spehro Pefhany

-----
http://www.mindspring.com/~benbradley

 

[Robert Baer]

Dirk Bruere at Neopax wrote:

Robert Baer wrote:

Dirk Bruere at Neopax wrote:

Ban wrote:

Dirk Bruere at Neopax wrote:

Dirk Bruere at Neopax wrote:


Hi

Looking for a low frequency signal generator chip that can put out a
sine wave varying in frequency from around 1Hz to 100Hz with
resolution around 0.01Hz
Any ideas for a cheap solution?




Thanks for all the suggestions guys, and I'll keep them in mind.
However, I have decided the best (ie cheapest and hassle free)method
is to use a program like Audacity on a PC to generate the signal I
need to high precision. I can then either output it directly from the
soundcard or convert it to MP3 and play it from a portable device.





I was just checking out with my old copy of CEP what happens when
you add a 1Hz and a 1.01Hz sinewave. nice beat frequency.
The question is if the soundcard can output 1Hz without attenuating
it a bit. You should at least increase the output capacitors, so the
lower frequency -3dB point is 0.1Hz or less.




Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff.
One solution is to create a stereo track with one channel at f1 and
the other at f2 where f1>f2, feed these into a summing op-amp and
follow by some simple filtering to get f1-f2.

That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.


Summing them won't work?

Harmonics are also generated.


The f1+f2 can be filtered out. It should be easy if f1,f2 >> f1-f2

Summing signal one and signal 2 in an operational amplifier cannot

create other frequencies, assuming the opamp is working in the linear
region.
Like i said, you need a non-linear device.

 

[Dirk Bruere at Neopax]

Rich Grise wrote:

On Thu, 28 Jul 2005 11:31:29 +0100, Dirk Bruere at Neopax wrote:

Robert Baer wrote:

Dirk Bruere at Neopax wrote:


Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff. One solution is
to create a stereo track with one channel at f1 and the other at f2
where f1>f2, feed these into a summing op-amp and follow by some simple
filtering to get f1-f2.


That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.

Summing them won't work?


No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

So what *does* come out if I put f1, f2 into a summing (or difference) op amp?

--
Dirk

The Consensus:-
The political party for the new millenium
http://www.theconsensus.org

 

[Don Bowey]

On 8/4/05 8:21 PM, in article 3lg49dF1257l1U1@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:

Rich Grise wrote:

On Thu, 28 Jul 2005 11:31:29 +0100, Dirk Bruere at Neopax wrote:

Robert Baer wrote:

Dirk Bruere at Neopax wrote:


Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff. One solution is
to create a stereo track with one channel at f1 and the other at f2
where f1>f2, feed these into a summing op-amp and follow by some simple
filtering to get f1-f2.


That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.

Summing them won't work?


No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

So what *does* come out if I put f1, f2 into a summing (or difference) op amp?

f1 and f2.

 

[Dirk Bruere at Neopax]

Don Bowey wrote:

On 8/4/05 8:21 PM, in article 3lg49dF1257l1U1@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:


Rich Grise wrote:


On Thu, 28 Jul 2005 11:31:29 +0100, Dirk Bruere at Neopax wrote:


Robert Baer wrote:


Dirk Bruere at Neopax wrote:



Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff. One solution is
to create a stereo track with one channel at f1 and the other at f2
where f1>f2, feed these into a summing op-amp and follow by some simple
filtering to get f1-f2.


That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.

Summing them won't work?


No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

So what *does* come out if I put f1, f2 into a summing (or difference) op amp?


f1 and f2.

So two sine waves go in, and from a single o/p comes...?
If I add two sine waves, f1 f1 in Audacity I get a signal amplitude modulated
100% at f1-f2

Why would a summing amp be different?

--
Dirk

The Consensus:-
The political party for the new millenium
http://www.theconsensus.org

 

[Ban]

Dirk Bruere at Neopax wrote:

Summing them won't work?


No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?


f1 and f2.

So two sine waves go in, and from a single o/p comes...?
If I add two sine waves, f1 f1 in Audacity I get a signal amplitude
modulated 100% at f1-f2

Why would a summing amp be different?

Dirk, I would advise you to make a frequency analysis in Audacity and then
tell us how high the level is at the beat frequency.

--
ciao Ban
Bordighera, Italy

 

[Don Bowey]

On 8/4/05 9:02 PM, in article 3lg6nfF12flt2U1@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:

Don Bowey wrote:

On 8/4/05 8:21 PM, in article 3lg49dF1257l1U1@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:


Rich Grise wrote:


On Thu, 28 Jul 2005 11:31:29 +0100, Dirk Bruere at Neopax wrote:


Robert Baer wrote:


Dirk Bruere at Neopax wrote:



Well, as far as I know Audacity can create a tone at a very precise
frequency.
However, you are correct about playing very LF stuff. One solution is
to create a stereo track with one channel at f1 and the other at f2
where f1>f2, feed these into a summing op-amp and follow by some simple
filtering to get f1-f2.


That will not work.
In order to get (cerate, or generate) a sum or difference of
frequencies, one needs to use those signals to drive a nonlinear device.

Summing them won't work?


No. You have to modulate one with the other, or multiply them with
an analog multiplier. In fact, I've heard of an analog multiplier
that's optimized for balanced modulator use.

So what *does* come out if I put f1, f2 into a summing (or difference) op
amp?


f1 and f2.

So two sine waves go in, and from a single o/p comes...?
If I add two sine waves, f1 f1 in Audacity I get a signal amplitude modulated
100% at f1-f2

Why would a summing amp be different?

You are not going to get a amplitude modulate signal by putting two signals
through a linear device. You need a non-linear element to "multiply" the
signals.

 

[Tony Williams]

In article <3lg49dF1257l1U1@individual.net>,
Dirk Bruere at Neopax <dirk@neopax.com> wrote:

So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?

SinA + SinB = 2 * Sin(A+B)/2 * Cos(A-B)/2.

Which I think on the scope should look like a sinewave at
a frequency of (A+B)/2, amplitude-modulated at (A-B)2.

OTOH.

SinA * SinB = 0.5 * [ Cos(A-B) - Cos(A+B) ].

Which produces separate sum and difference frequencies.

--
Tony Williams.

 

[KoKlust]

"Tony Williams" <tonyw@ledelec.demon.co.uk> wrote in message
news:4d9587c1b1tonyw@ledelec.demon.co.uk...

In article <3lg49dF1257l1U1@individual.net>,
Dirk Bruere at Neopax <dirk@neopax.com> wrote:

So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?

Out the opamp comes V1 = cos(f1*t) + cos(f2*t) .

Periodically with f2-f1 (or f1-f2), the two cosines will interfere
constructively or destructively. So the ENVELOPE of the signal does contain
f2-f1.

STILL if you would look on a spectrum analyser you see only TWO different
frequencies.

So the envelope can be detected by rectifying and low-pass filtering Vout,
to get the V2 = a*cos((f2-f1)*t) signal. (AM-detection). A very simple
circuit that does this would be:

V1 --|>|---+----+---- V2
| |
| | ---
R | | --- C
| |
GND --------+----+---- GND

I guess that RC must be chosen somewhere around 1/(2*pi*fh), where fh is the
highest of f1 and f2.


In short: The nonlinear element (diode in this case) is essential, and using
only an opamp will not bring you the difference frequency.

 

[Dirk Bruere at Neopax]

KoKlust wrote:

"Tony Williams" <tonyw@ledelec.demon.co.uk> wrote in message
news:4d9587c1b1tonyw@ledelec.demon.co.uk...

In article <3lg49dF1257l1U1@individual.net>,
Dirk Bruere at Neopax <dirk@neopax.com> wrote:


So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?


Out the opamp comes V1 = cos(f1*t) + cos(f2*t) .

Periodically with f2-f1 (or f1-f2), the two cosines will interfere
constructively or destructively. So the ENVELOPE of the signal does contain
f2-f1.

STILL if you would look on a spectrum analyser you see only TWO different
frequencies.

So the envelope can be detected by rectifying and low-pass filtering Vout,
to get the V2 = a*cos((f2-f1)*t) signal. (AM-detection). A very simple
circuit that does this would be:

V1 --|>|---+----+---- V2
| |
| | ---
R | | --- C
| |
GND --------+----+---- GND

I guess that RC must be chosen somewhere around 1/(2*pi*fh), where fh is the
highest of f1 and f2.


In short: The nonlinear element (diode in this case) is essential, and using
only an opamp will not bring you the difference frequency.

Well, yes.
That's what I was talking about originally (ie adding post processing) but kept
being told it wasn't possible. Seems everyone looks at the letter of the law and
not the spirit.

Thanks
Dirk

The Consensus:-
The political party for the new millenium
http://www.theconsensus.org

 

[Don Bowey]

On 8/5/05 8:36 AM, in article 3lhfcqF12qg3pU2@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:

KoKlust wrote:
"Tony Williams" <tonyw@ledelec.demon.co.uk> wrote in message
news:4d9587c1b1tonyw@ledelec.demon.co.uk...

In article <3lg49dF1257l1U1@individual.net>,
Dirk Bruere at Neopax <dirk@neopax.com> wrote:


So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?


Out the opamp comes V1 = cos(f1*t) + cos(f2*t) .

Periodically with f2-f1 (or f1-f2), the two cosines will interfere
constructively or destructively. So the ENVELOPE of the signal does contain
f2-f1.

STILL if you would look on a spectrum analyser you see only TWO different
frequencies.

So the envelope can be detected by rectifying and low-pass filtering Vout,
to get the V2 = a*cos((f2-f1)*t) signal. (AM-detection). A very simple
circuit that does this would be:

V1 --|>|---+----+---- V2
| |
| | ---
R | | --- C
| |
GND --------+----+---- GND

I guess that RC must be chosen somewhere around 1/(2*pi*fh), where fh is the
highest of f1 and f2.


In short: The nonlinear element (diode in this case) is essential, and using
only an opamp will not bring you the difference frequency.

Well, yes.
That's what I was talking about originally (ie adding post processing) but
kept
being told it wasn't possible. Seems everyone looks at the letter of the law
and
not the spirit.

Bull!

That is not what you said or even implied in previous posts.

Each time someone fed you info about what you need to do to obtain a sum and
difference signal, you argued. What we all told you was simple facts about
amplitude modulation. Law? Yes. The only post processing you talked about
was "simple filtering to get f1-f2." That won't work no matter what is the
assumed "spirit" of what you want to do.

 

[KoKlust]

In short: The nonlinear element (diode in this case) is essential, and
using only an opamp will not bring you the difference frequency.

Well, yes.
That's what I was talking about originally (ie adding post processing) but
kept being told it wasn't possible. Seems everyone looks at the letter of
the law and not the spirit.

Thanks
Dirk

You're welcome Dirk.

 

[Dirk Bruere at Neopax]

Don Bowey wrote:

On 8/5/05 8:36 AM, in article 3lhfcqF12qg3pU2@individual.net, "Dirk Bruere
at Neopax" <dirk@neopax.com> wrote:


KoKlust wrote:

"Tony Williams" <tonyw@ledelec.demon.co.uk> wrote in message
news:4d9587c1b1tonyw@ledelec.demon.co.uk...


In article <3lg49dF1257l1U1@individual.net>,
Dirk Bruere at Neopax <dirk@neopax.com> wrote:



So what *does* come out if I put f1, f2 into a summing (or
difference) op amp?


Out the opamp comes V1 = cos(f1*t) + cos(f2*t) .

Periodically with f2-f1 (or f1-f2), the two cosines will interfere
constructively or destructively. So the ENVELOPE of the signal does contain
f2-f1.

STILL if you would look on a spectrum analyser you see only TWO different
frequencies.

So the envelope can be detected by rectifying and low-pass filtering Vout,
to get the V2 = a*cos((f2-f1)*t) signal. (AM-detection). A very simple
circuit that does this would be:

V1 --|>|---+----+---- V2
| |
| | ---
R | | --- C
| |
GND --------+----+---- GND

I guess that RC must be chosen somewhere around 1/(2*pi*fh), where fh is the
highest of f1 and f2.


In short: The nonlinear element (diode in this case) is essential, and using
only an opamp will not bring you the difference frequency.

Well, yes.
That's what I was talking about originally (ie adding post processing) but
kept
being told it wasn't possible. Seems everyone looks at the letter of the law
and
not the spirit.


Bull!

That is not what you said or even implied in previous posts.

Each time someone fed you info about what you need to do to obtain a sum and
difference signal, you argued. What we all told you was simple facts about
amplitude modulation. Law? Yes. The only post processing you talked about
was "simple filtering to get f1-f2." That won't work no matter what is the
assumed "spirit" of what you want to do.

The point being that in order to get what I want a summing op amp is essential
as part of the solution. I had that, and so I needed the rest of it. Which
others have kindly supplied.

--
Dirk

The Consensus:-
The political party for the new millenium
http://www.theconsensus.org