Maker Pro
Maker Pro

Active low-pass filter with gain

M

Michael

Jan 1, 1970
0
I'm using an LM741 as an active low-pass filter. Non-inverting input
with output capacitor feedback between two series resistors. This works
fine with unity gain (output tied back to inverting input). However, if
I try to increase the gain using a resistor divider to feedback to the
inverting input, I stop getting any output (however, if I remove the
inverting feedback completely, I seem to get an extremely large gain).

Any ideas on what's happening.

TIA.
 
N

Noway2

Jan 1, 1970
0
I'm using an LM741 as an active low-pass filter.  Non-inverting input
with output capacitor feedback between two series resistors.  This works
fine with unity gain (output tied back to inverting input).  However, if
I try to increase the gain using a resistor divider to feedback to the
inverting input, I stop getting any output (however, if I remove the
inverting feedback completely, I seem to get an extremely large gain).

Any ideas on what's happening.

TIA.

It is difficult to fully understand your circuit by your description,
especially the part about "output capacitor feedback between two
series resistors".

I would first suggest using a a standard filter topology, suck as a
Sallen Key, which will reduce the filter design to matching against a
set of normalized charts and then scaling for frequency and amplitude.

Another way to look at this is to remove the frequency selective
components (ie the cap), and concentrate on the amplifier portion
itself. A non-inverting amplifier configuration has a gain of 1 +
R_feedback / R_input. Analyze the circuit until you understand why.
The opamp will TRY to drive the output such that the voltage between
its input terminals is zero. This creates a 'virtual short' between
the inputs which will simplify your analysis. Once you have the
equation (transfer function) describing the amplifier, you can put the
capacitor back in the circuit and re-analyze. You will want to use
the LaPlace transform when performing your analysis to convert the
differential equations caused by the inclusion of the cap to a
frequency domain for use in filtering.
 
M

Michael Black

Jan 1, 1970
0
Noway2 said:
It is difficult to fully understand your circuit by your description,
especially the part about "output capacitor feedback between two
series resistors".

I would first suggest using a a standard filter topology, suck as a
Sallen Key, which will reduce the filter design to matching against a
set of normalized charts and then scaling for frequency and amplitude.
I think he is describing a Sallen-Key. The input signal goes to
the non-inverting input of the op-amp, with the input resistor broken
down into two. So you have to stages of an RC low pass filter, with
the capacitor of the first stage connected to the output of the op-amp
rather than ground.

He's trying to get gain from it by adjusting the gain of the op-amp,
which isn't going to happen. Likely his "no output" is because the
signal is swamped by oscillation, though that doesn't fully explain why
he sees output when he has no negative feedback (which of course would
give "near infinite" gain.

Michael
 
M

Michael

Jan 1, 1970
0
Michael said:
I think he is describing a Sallen-Key. The input signal goes to
the non-inverting input of the op-amp, with the input resistor broken
down into two. So you have to stages of an RC low pass filter, with
the capacitor of the first stage connected to the output of the op-amp
rather than ground.

He's trying to get gain from it by adjusting the gain of the op-amp,
which isn't going to happen. Likely his "no output" is because the
signal is swamped by oscillation, though that doesn't fully explain why
he sees output when he has no negative feedback (which of course would
give "near infinite" gain.

Michael
You're correct, it looks like a Sallen-Key filter (the R and C values
came from the output of Microchip's FilterLab). And the output with no
feedback sounds (and looks) like "near infinite" gain (severe clipping).
Is there a way to prevent the oscillation problem and get some
gain from this stage? I have been able to get some gain by taking the
output to another LM741 stage, but I assume I need to clean things up -
especially since it's on a breadboard (and not very neat either).
 
J

John Popelish

Jan 1, 1970
0
Michael said:
You're correct, it looks like a Sallen-Key filter (the R and C values
came from the output of Microchip's FilterLab). And the output with no
feedback sounds (and looks) like "near infinite" gain (severe clipping).
Is there a way to prevent the oscillation problem and get some
gain from this stage? I have been able to get some gain by taking the
output to another LM741 stage, but I assume I need to clean things up -
especially since it's on a breadboard (and not very neat either).

I'm surprised Microchip's FilterLab doesn't include the
ability to set the gain as you wish. Texas Instruments
FilterPro does.

http://focus.ti.com/docs/toolsw/folders/print/filterpro.html
 
B

Bob Masta

Jan 1, 1970
0
You're correct, it looks like a Sallen-Key filter (the R and C values
came from the output of Microchip's FilterLab). And the output with no
feedback sounds (and looks) like "near infinite" gain (severe clipping).
Is there a way to prevent the oscillation problem and get some
gain from this stage? I have been able to get some gain by taking the
output to another LM741 stage, but I assume I need to clean things up -
especially since it's on a breadboard (and not very neat either).

Without looking at the circuit I can't be sure, but note that the gain
of the filter is often used to determine Q. Your circuit description
sounds like a 2nd order low-pass, except you haven't mentioned
a second capacitor to ground. Usually, if you increase gain you
increase Q, and you will soon get into oscillation... not to mention
putting a huge peak at the corner frequency.

Check out Don Lancaster's "Active Filter Cookbook" for some
reliable circuits. (And note that the 741 is probably not a good
choice any more, though it should be fine at low frequencies
if low noise is not important. Consider the LF351 or TL071
series, for example, if you want drop-in parts with much better
performance.)

Best regards,


Bob Masta

DAQARTA v3.50
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, FREE Signal Generator
Science with your sound card!
 
M

Michael

Jan 1, 1970
0
Bob said:
Without looking at the circuit I can't be sure, but note that the gain
of the filter is often used to determine Q. Your circuit description
sounds like a 2nd order low-pass, except you haven't mentioned
a second capacitor to ground. Usually, if you increase gain you
increase Q, and you will soon get into oscillation... not to mention
putting a huge peak at the corner frequency.

Check out Don Lancaster's "Active Filter Cookbook" for some
reliable circuits. (And note that the 741 is probably not a good
choice any more, though it should be fine at low frequencies
if low noise is not important. Consider the LF351 or TL071
series, for example, if you want drop-in parts with much better
performance.)

Best regards,


Bob Masta

DAQARTA v3.50
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, FREE Signal Generator
Science with your sound card!
Thanks for the book reference. The library has a copy I can check out.
 
J

John Popelish

Jan 1, 1970
0
Michael said:
John Popelish wrote:
Thanks for the link. The program with Sallen-Key selected shows exactly
what I'm trying to do, so I'm assuming oscillation is preventing the gain.

You have mentioned only one capacitor in your filter. The
minimum filter this program generates uses two. and you
should also have a small capacitor between each of the
supply pins and ground, across your opamp, to stabilize it.
 
M

Michael

Jan 1, 1970
0
John said:
You have mentioned only one capacitor in your filter. The minimum
filter this program generates uses two. and you should also have a
small capacitor between each of the supply pins and ground, across your
opamp, to stabilize it.
I have the cap on the non-inverting input, but not those on the supply.
Recently encountered this article which mentioned them:

www.jensen-transformers.com/an/an001.pdf

Will apply caps and see what results.

Thanks again.
 
Top