Wanting a circuit to find the average of a fluctuating DC si

J

jalbers@bsu.edu

Guest
I am looking for a circuit that will give me the average of a
fluctuating DC signal. I am using the term average to roughly mean a
value in the middle but doens not have to be exact.

A few years ago I was told by some people in this group to simply use
a low pass filter. At that time I did not have the equipment, time,
or knowledge of filters to mess with it. I have started to play
around with different types of filters...

I have done a little bit of experimenting using a scope, signal
generator, and a simple low pass RC filter but can't get it to work to
find the average of the signal.

I am ultimately interested in averaging the fluctuating DC signal from
a homemade EDM machine which is a basically a big DC relaxation
oscillator. For various reasons, I do not want to use the actual EDM
machine for these experiments. I have been using a function generator
and also a NE-2 relaxation oscillator to somewhat simulate the signal
that the EDM machine is going to be putting off. I have experimented
with sine, triangle, square, and relaxation oscillator wave forms with
enough DC offset to push the entire waveform completely up above the x-
axis on the scope. To no surprize, increasing the frequency begins to
attenuate the signal and the output of the low pass filter starts to
hug the x-axis. The capacitor is removing the DC component and
attenuating the AC part of the signal.

I don't think that a LC low pass filter is going to work the way I
want it to either. It will not remove the DC component but the
waveform will still hug the x-axis.

I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.

Any help would be greatly appreciated. Thanks
 
On Wed, 23 Jul 2008 08:14:57 -0700 (PDT), "jalbers@bsu.edu"
<jalbers@bsu.edu> wrote:

I am looking for a circuit that will give me the average of a
fluctuating DC signal. I am using the term average to roughly mean a
value in the middle but doens not have to be exact.

A few years ago I was told by some people in this group to simply use
a low pass filter. At that time I did not have the equipment, time,
or knowledge of filters to mess with it. I have started to play
around with different types of filters...

I have done a little bit of experimenting using a scope, signal
generator, and a simple low pass RC filter but can't get it to work to
find the average of the signal.

I am ultimately interested in averaging the fluctuating DC signal from
a homemade EDM machine which is a basically a big DC relaxation
oscillator. For various reasons, I do not want to use the actual EDM
machine for these experiments. I have been using a function generator
and also a NE-2 relaxation oscillator to somewhat simulate the signal
that the EDM machine is going to be putting off. I have experimented
with sine, triangle, square, and relaxation oscillator wave forms with
enough DC offset to push the entire waveform completely up above the x-
axis on the scope. To no surprize, increasing the frequency begins to
attenuate the signal and the output of the low pass filter starts to
hug the x-axis. The capacitor is removing the DC component and
attenuating the AC part of the signal.

I don't think that a LC low pass filter is going to work the way I
want it to either. It will not remove the DC component but the
waveform will still hug the x-axis.

I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.

Any help would be greatly appreciated. Thanks
Click to expand...
First, define "average."

John
 
jalbers@bsu.edu wrote:
(snip)
I am ultimately interested in averaging the fluctuating DC signal from
a homemade EDM machine which is a basically a big DC relaxation
oscillator. For various reasons, I do not want to use the actual EDM
machine for these experiments. I have been using a function generator
and also a NE-2 relaxation oscillator to somewhat simulate the signal
that the EDM machine is going to be putting off. I have experimented
with sine, triangle, square, and relaxation oscillator wave forms with
enough DC offset to push the entire waveform completely up above the x-
axis on the scope. To no surprize, increasing the frequency begins to
attenuate the signal and the output of the low pass filter starts to
hug the x-axis. The capacitor is removing the DC component and
attenuating the AC part of the signal.
(snip)
I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.
Click to expand...
Swap the positions of the R and C in your filter. The R
should connect to the oscillator signal and the C should
connect to ground.

--
Regards,

John Popelish
 
<jalbers@bsu.edu> wrote in message
news:841c4ffa-5ab4-4bd7-bdd6-0e9e59dd320d@j22g2000hsf.googlegroups.com...
I am looking for a circuit that will give me the average of a
fluctuating DC signal. I am using the term average to roughly mean a
value in the middle but doens not have to be exact.

A few years ago I was told by some people in this group to simply use
a low pass filter. At that time I did not have the equipment, time,
or knowledge of filters to mess with it. I have started to play
around with different types of filters...

I have done a little bit of experimenting using a scope, signal
generator, and a simple low pass RC filter but can't get it to work to
find the average of the signal.

I am ultimately interested in averaging the fluctuating DC signal from
a homemade EDM machine which is a basically a big DC relaxation
oscillator. For various reasons, I do not want to use the actual EDM
machine for these experiments. I have been using a function generator
and also a NE-2 relaxation oscillator to somewhat simulate the signal
that the EDM machine is going to be putting off. I have experimented
with sine, triangle, square, and relaxation oscillator wave forms with
enough DC offset to push the entire waveform completely up above the x-
axis on the scope. To no surprize, increasing the frequency begins to
attenuate the signal and the output of the low pass filter starts to
hug the x-axis. The capacitor is removing the DC component and
attenuating the AC part of the signal.

I don't think that a LC low pass filter is going to work the way I
want it to either. It will not remove the DC component but the
waveform will still hug the x-axis.

I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.

Any help would be greatly appreciated. Thanks
Click to expand...

As Mr Larken points out, you need to define your concept of 'average'.

You need to define

1) The period over which you want the average taken
2) The frequencies your input signal contains
3) How you want to read the average value obtained (ie, what kind of a
circuit will use the output).
4) How much the 'averaging' circuit may affect the input.

That is all I can think of, but Mr Larken may have more ideas. Once the
requirements are defined, it'll be easier to help you.

Regards,
Bob Monsen
 
On Jul 23, 12:48 pm, John Popelish <jpopel...@rica.net> wrote:
jalb...@bsu.edu wrote:

(snip)

I am ultimately interested in averaging the fluctuating DC signal from
a homemade EDM machine which is a basically a big DC relaxation
oscillator.   For various reasons, I do not want to use the actual EDM
machine for these experiments.  I have been using a function generator
and also a NE-2 relaxation oscillator to somewhat simulate the signal
that the EDM machine is going to be putting off.  I have experimented
with sine, triangle, square, and relaxation oscillator wave forms with
enough DC offset to push the entire waveform completely up above the x-
axis on the scope.  To no surprize, increasing the frequency begins to
attenuate the signal and the output of the low pass filter starts to
hug the x-axis.  The capacitor is removing the DC component and
attenuating the AC part of the signal.
(snip)
I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.

Swap the positions of the R and C in your filter.  The R
should connect to the oscillator signal and the C should
connect to ground.

--
Regards,

John Popelish
Click to expand...
Thanks for figuring out from the info I had given that I had the
connections switched around. The circuit seems to work correctly, as
the frequency of the the signal produced by the function generator
increases the voltage across C decreases but is centered around the DC
offset and the voltage across R increases but is centered around zero.

I am however having a little trouble using this same RC low pass
filter on the NE-2 relaxation oscillator circuit. The relaxation
oscillator is producing a nice sawtooth wave centered around 70 VDC on
the scope but when I connect the output of the relaxation oscillator
to the RC low pass filter and chage the frequency, I am not seeing the
voltage across C (sawtooth centered around 70 VDC) decreasing in
amplitude and the components in the low pass filter are srewing with
the frequency of the relaction oscillator.

The relaxation oscillator I am using consists of a 1M resistor wired
in series with NE-2 wired parallel to a capacitor and all of this is
conned to a 150VDC supply. I am using a .47uF, 11nF, and 1.9nF to
generate 9Hz, 333Hz, and 2500Hz signals. The RC filter consists of a
1K resistor and .1 uF capacitor. What size resistor and capacitor
shoud I be using in the RC low pass filter to be able to observe the
wave form decreasing in amplitude around the 70VDC offset?
 
jalbers@bsu.edu wrote:

Thanks for figuring out from the info I had given that I had the
connections switched around. The circuit seems to work correctly, as
the frequency of the the signal produced by the function generator
increases the voltage across C decreases but is centered around the DC
offset and the voltage across R increases but is centered around zero.

I am however having a little trouble using this same RC low pass
filter on the NE-2 relaxation oscillator circuit. The relaxation
oscillator is producing a nice sawtooth wave centered around 70 VDC on
the scope but when I connect the output of the relaxation oscillator
to the RC low pass filter and chage the frequency, I am not seeing the
voltage across C (sawtooth centered around 70 VDC) decreasing in
amplitude and the components in the low pass filter are srewing with
the frequency of the relaction oscillator.

The relaxation oscillator I am using consists of a 1M resistor wired
in series with NE-2 wired parallel to a capacitor and all of this is
conned to a 150VDC supply. I am using a .47uF, 11nF, and 1.9nF to
generate 9Hz, 333Hz, and 2500Hz signals. The RC filter consists of a
1K resistor and .1 uF capacitor. What size resistor and capacitor
shoud I be using in the RC low pass filter to be able to observe the
wave form decreasing in amplitude around the 70VDC offset?
Click to expand...
The problem is that the oscillator is a very low current
circuit (powered through a 1 meg resistor), so the filter
must use even less current in order to not badly distort the
oscillator's operation. For instance, the R in the RC
filter may have to be something in the 2 to 10meg range.

Expect the scope or meter load on the capacitor to lower the
average DC voltage, since the resistance of those loads will
be pretty low compared to the filter resistance.

--
Regards,

John Popelish
 
On Jul 23, 11:14 am, "jalb...@bsu.edu" <jalb...@bsu.edu> wrote:
I am looking for a circuit that will give me the average of a
fluctuating DC signal. ... from
a homemade EDM machine
Click to expand...
The gap voltage peak and gap voltage average have a useful
relationship,
or the average current might be useful, so let's consider those
first.

Current can be related simply to the power supply input current; just
measure the input current if you want that (the power supply
filter will keep fluctuations relatively slow).

The gap voltage rings/oscillates and has fast-slewing character, so it
isn't
nice to attach a long wire to this; you will want to have a filter
near
the spark gap that attenuates RF (maybe .01 uF/100 kohm resistor),
as well as a slow filter (two seconds time constant?) using an op
amp as an integrator. Analog (moving-needle) meters are possibly
slow enough to just connect directly, using the inertia of the moving
assembly.

It's possible, too, that your concern is with the area under a single
pulse
X = integral(gap_voltage * dT| T= 0 to T = end-of-pulse)
(irrespective of time separation of pulses) which requires a
resettable
circuit. First, discharge a capacitor. Then, set a timer to open a
gate
when a pulse is detected, and integrate (onto the capacitor) current
proportional to the gap voltage until the timer closes the gate. Then
display the capacitor's voltage on a meter, using a sample/hold
circuit. It sounds complicated, but a dual op amp and quad analog
switch
will do the job.
 
On Wed, 23 Jul 2008 08:14:57 -0700, jalbers@bsu.edu wrote:

I am looking for something that will attenuate the signal as the
frequency increases but still keep the DC component of the signal.

Any help would be greatly appreciated.
Click to expand...
Signal in >-----[R]-----+--------> DC out
|
[C]
|
GND

Determine the R and C values by the formula f = 1/(2 * PI * SQRT(R * C))
approx; then tweak until you get the result you want. (a higher-value
resistor OR a higher-value capacitor (or both) will decrease the cutoff
frequency.

Have Fun!
Rich
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top