Input signal distortion.. capacitive coupling?

[MRW]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

For my VCC, I used a 3V output from a DC power supply. For VCC/2, I
used 1.5V from another output of the DC power supply. My input signal
came from a soundcard output. I used a program called DAQGen to
generate the 1kHz tone.

I measured my component values and they came fairly close to the design
values. However, when I connect the sound card output to the filter
input, my 1 kHz sine wave gets distorted. You can see it in the image
link. I am wondering what causes this.

I read in the Art of Electronics book that if I see some spikes on my
sinewave, then it is an indication that I may have some unintentional
capacitive coupling from a square wave. But I don't think I have any
square wave signals. Also, the spikes seem to be occurring
approximately every 450us (~2222 Hz).

What would cause my signal to get this kind of distortion when I plug
it into the filter input?

Thanks!

 

[Michael Black]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

I don't have a clue, but "distortion caused by capacitors" is often the
realm of people who spend Big Bucks on audio equipment, and then pay
good money to replace the capacitors because they believe the existing
capacitors are causing distortion.

Take note that it's well known that the 324 is not a good op-amp. It's
pretty well known that you should put a resistor from the output to
ground to help to get rid of the common distortion problem. I don't
remember the full details, but they should be easily available


Michael

 

[Andrew Holme]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

For my VCC, I used a 3V output from a DC power supply. For VCC/2, I
used 1.5V from another output of the DC power supply. My input signal
came from a soundcard output. I used a program called DAQGen to
generate the 1kHz tone.

I measured my component values and they came fairly close to the design
values. However, when I connect the sound card output to the filter
input, my 1 kHz sine wave gets distorted. You can see it in the image
link. I am wondering what causes this.

I read in the Art of Electronics book that if I see some spikes on my
sinewave, then it is an indication that I may have some unintentional
capacitive coupling from a square wave. But I don't think I have any
square wave signals. Also, the spikes seem to be occurring
approximately every 450us (~2222 Hz).

What would cause my signal to get this kind of distortion when I plug
it into the filter input?

Thanks!

Is the'scope input connected to the souncard output (filter input) in all
three cases? Try running the op-amp from a slightly higher supply voltage.
3V is a bit low.

 

[Fred Bartoli]

Michael Black a écrit :

I don't have a clue, but "distortion caused by capacitors" is often the
realm of people who spend Big Bucks on audio equipment, and then pay
good money to replace the capacitors because they believe the existing
capacitors are causing distortion.

Take note that it's well known that the 324 is not a good op-amp. It's
pretty well known that you should put a resistor from the output to
ground to help to get rid of the common distortion problem. I don't
remember the full details, but they should be easily available

LM324 is all but a 3V opamp, despite what's said in the datasheet.
At least you can't use it with its bias point only 1.5V below the
positive rail (look input CM range and output limits).

If you want 3V supply, then use an adequate rail to rail opamp.

 

[John Popelish]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

For my VCC, I used a 3V output from a DC power supply. For VCC/2, I
used 1.5V from another output of the DC power supply. My input signal
came from a soundcard output. I used a program called DAQGen to
generate the 1kHz tone.

I measured my component values and they came fairly close to the design
values. However, when I connect the sound card output to the filter
input, my 1 kHz sine wave gets distorted. You can see it in the image
link. I am wondering what causes this.

I read in the Art of Electronics book that if I see some spikes on my
sinewave, then it is an indication that I may have some unintentional
capacitive coupling from a square wave. But I don't think I have any
square wave signals. Also, the spikes seem to be occurring
approximately every 450us (~2222 Hz).

What would cause my signal to get this kind of distortion when I plug
it into the filter input?

Some comments:

The LM324 is a very low current opamp, but you pay for that
low supply current with a class B output stage that has a
dead spot, every time the current reverses. Adding an
output load resistor to a supply rail, that keeps the
current from reversing adds to the supply current but
improves this cross over distortion.

The LM324 just barely operates on 3 volts. I suggest you do
your experiments with a pair of 9 volt batteries in series
as a supply, with the center point as signal ground. After
you get everything working, you can try reducing the total
supply voltage as low as about 5 volts, unless you get a
better opamp.

 

[Eeyore]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

LMAO.

LM324 !

What do you expect. Use something half decent at least !

Graham

 

[Eeyore]

I don't have a clue, but "distortion caused by capacitors" is often the
realm of people who spend Big Bucks on audio equipment, and then pay
good money to replace the capacitors because they believe the existing
capacitors are causing distortion.

Take note that it's well known that the 324 is not a good op-amp. It's
pretty well known that you should put a resistor from the output to
ground to help to get rid of the common distortion problem. I don't
remember the full details, but they should be easily available

You mean from output to negative supply actually but a decent op-amp would be a far better
bet.

Graham

 

[Eeyore]

Try running the op-amp from a slightly higher supply voltage.
3V is a bit low.

Goodness, I missed that !

A bit low is putting it mildly.

Graham

 

[ehsjr]

Hello again!

Let me put some links before I start my question.

Image:
http://img79.imageshack.us/img79/9045/oscopeoutfilterxg7.jpg

Texas Instruments filter wizard:
http://www-k.ext.ti.com/SRVS/Data/ti/KnowledgeBases/analog/document/faqs/fwizard.htm

I decided to replicate the 2-pole high pass filter given by the TI
wizard (also seen in the image link). I picked a cut off frequency of
20 Hz and C = 1 uF. I calculated the R to be about 7958 ohms, so my R1
and R2 values were: 3716 ohms and 16791 ohms. I used one of the opamps
of the LMT324N quad opamp IC; I couldn't find the datasheet for this so
I used the LM324's.

For my VCC, I used a 3V output from a DC power supply. For VCC/2, I
used 1.5V from another output of the DC power supply. My input signal
came from a soundcard output. I used a program called DAQGen to
generate the 1kHz tone.

I measured my component values and they came fairly close to the design
values. However, when I connect the sound card output to the filter
input, my 1 kHz sine wave gets distorted. You can see it in the image
link. I am wondering what causes this.

I read in the Art of Electronics book that if I see some spikes on my
sinewave, then it is an indication that I may have some unintentional
capacitive coupling from a square wave. But I don't think I have any
square wave signals. Also, the spikes seem to be occurring
approximately every 450us (~2222 Hz).

What would cause my signal to get this kind of distortion when I plug
it into the filter input?

Thanks!

Not adding to what's already been said or addressing
the problem, just a question: what are you using for
the oscilloscope?

Ed

 

[MRW]

Take note that it's well known that the 324 is not a good op-amp. It's
pretty well known that you should put a resistor from the output to
ground to help to get rid of the common distortion problem. I don't
remember the full details, but they should be easily available

I haven't been around with electronics long enough to know this.
Everyone's got to start somewhere.

 

[MRW]

Is the'scope input connected to the souncard output (filter input) in all
three cases? Try running the op-amp from a slightly higher supply voltage.
3V is a bit low.

Yeah, I didn't change the location of the scope probe at all. As for
the supply voltage, I just went by with what the datasheet said. It
said that I could operate it with 3V in single supply mode.

 

[MRW]

The LM324 is a very low current opamp, but you pay for that
low supply current with a class B output stage that has a
dead spot, every time the current reverses. Adding an
output load resistor to a supply rail, that keeps the
current from reversing adds to the supply current but
improves this cross over distortion.

I'm still trying to understand this statement. Sorry.

The LM324 just barely operates on 3 volts. I suggest you do
your experiments with a pair of 9 volt batteries in series
as a supply, with the center point as signal ground. After
you get everything working, you can try reducing the total
supply voltage as low as about 5 volts, unless you get a
better opamp.

Operate the opamp in a split supply mode. Did I read this correctly?
I'll try this out, but ultimately, I'd like to try a single supply
operation.

Thanks!

 

[MRW]

LMAO.

LM324 !

What do you expect. Use something half decent at least !

Graham

I didn't know what to expect.

 

[MRW]

Not adding to what's already been said or addressing
the problem, just a question: what are you using for
the oscilloscope?

Ed

Hi Ed,

I'm using one of those Instek digital storage scopes. I'm still looking
if I can find an old analog scope for audio. Some of the ones that I
found are too much for my budget.

 

[Kingcosmos]

Operate the opamp in a split supply mode. Did I read this correctly?
I'll try this out, but ultimately, I'd like to try a single supply
operation.

Yep. I will tell you what my senior apps told me about single supply
op-amps. The world was happy when op-amps ran on +/-15V supplies.
Then some guy came along and said, "let us use +/-12V supplies" and
things still ran ok. Then some asshole came along and said, "Hey, let
us use a +5V supply" and the world has struggled ever since.

But Mr. Popelish is right. I have also seen some of TI's power op-amps
exhibit similar properties. You need to burn more quiescent current to
help eliminate cross over distortion inherent in the output stage.

 

[John Popelish]

I'm still trying to understand this statement. Sorry.

It is pretty simple, in concept. One glob of transistors
constitute the "pull up" system (that provides a current
path from the positive supply rail to the load) and another
glob of transistors constitute the "pull down" system (that
provides a current path from the load to the negative supply
rail). If both these systems are partly on, when no load
current is needed, this produces a current path from the
positive supply rail to the negative rail, draining power
from the supply, continuously. This is called class AB.

But it allows the amplifier to switch from pull up to pull
down, very smoothly, by just increasing the pull down
current while simultaneously reducing the pull up current
(or vice versa). The LM324 was optimized for very low
drain, battery operation, so instead of class AB, they set
it up so that, at idle, neither pull up or pull down is in
operation, so the output stage draws no current. But it
takes a little time to get the pull up or pull down system
operating, once either is needed. This dead time represents
a chunk of the output waveform that is malformed, and is
called cross over distortion (when the output current
crosses over from positive to negative, or vice versa).

Operate the opamp in a split supply mode. Did I read this correctly?

Yes. It simplifies lots of aspects of most circuits. Once
you understand these fewer aspects, then you add the details
that let the system operate from one supply.

I'll try this out, but ultimately, I'd like to try a single supply
operation.

I am talking about your education, not your final circuit.

 

[MRW]

It is pretty simple, in concept. One glob of transistors
constitute the "pull up" system (that provides a current
path from the positive supply rail to the load) and another
glob of transistors constitute the "pull down" system (that
provides a current path from the load to the negative supply
rail). If both these systems are partly on, when no load
current is needed, this produces a current path from the
positive supply rail to the negative rail, draining power
from the supply, continuously. This is called class AB.

Thanks! This is much clearer.

I am talking about your education, not your final circuit.

Yes, I'm working on that, too. I really haven't found the right mentor
to guide me with this electronics stuff, so I resort to reading books
and asking my questions here. I also try to read the magazines like the
IEEE Spectrum, Nuts & Volts, etc., but I haven't quite got all the
lingo set. A friend of a friend once told me that "once you know how to
deal with high frequency stuff, then the equations used for them will
also apply to the lower frequency counterparts." So I've decided to
also look into Smith Charts and S-Parameters to start things off. I
haven't quite gotten the big picture, yet, but I'm starting to put
puzzle pieces together.

 

[Eeyore]

Yeah, I didn't change the location of the scope probe at all. As for
the supply voltage, I just went by with what the datasheet said. It
said that I could operate it with 3V in single supply mode.

Except you're not using it in true single-supply mode. You're biasing the inputs
to 1/2 Vcc.

Graham

 

[Eeyore]

I didn't know what to expect.

Avoid using them at all costs for audio especially. Few worse op-amps have ever
been made.

Graham

 

[John Popelish]

Avoid using them at all costs for audio especially. Few worse op-amps have ever
been made.

I have to agree for any quality audio application (they may
be fine for an intercom). But for a vast number of low
frequency applications, their DC accuracy, power
consumption, availability and cost are hard to beat.