MAX186 and op-amps

[V8TR4]

Hi,

I made a small pic based project a while back that takes the outputs from
some off the shelf pressure sensors with a linear 0-10vdc output signal. The
output of the sensor goes thru a voltage devider into a MAX186 ADC, and then
into my PIC.

I have had some oscillation problems. The signals will oscillate up and down
and for some reason the signal coming in is not interpeted as being the
linear signal I expected. I added some caps around the board and the problem
became worse.

I looked into it and I see that I could use some op-amps to clean the signal
up. The MAX186 is a 0-4.096 ADC. I wanted to use the LM324, but I read that
it will give me problems due to the operating voltage. So it looks like the
741 might be a better choice. I am using all 8 channels of the ADC so eight
741's will take up a lot of board space.

Anyone have some hints and tips for what I am trying to do? What is the
cause of these oscillations and non linear numbers?

thanks in advance,
Oliver

 

[Walter Harley]

"V8TR4" <_v8tr4_@don't_bother_me_with_spam.com> wrote in message
newsuVib.13470$YO5.6092568@news3.news.adelphia.net...

I have had some oscillation problems. The signals will oscillate up and
down
and for some reason the signal coming in is not interpeted as being the
linear signal I expected. I added some caps around the board and the
problem
became worse.

It is possible that when the voltage from the sensors increases, it
increases current draw somewhere in your system, which causes the supply
voltage (to the whole thing, or just to the sensors) to sag, which causes
the voltage output to drop proportionally, etc....

What frequency of oscillation are you talking about?

I looked into it and I see that I could use some op-amps to clean the
signal
up. The MAX186 is a 0-4.096 ADC. I wanted to use the LM324, but I read
that
it will give me problems due to the operating voltage. So it looks like
the
741 might be a better choice. I am using all 8 channels of the ADC so
eight
741's will take up a lot of board space.

I'm not at all sure what you mean by "use some op-amps to clean the signal
up." Can you be more specific?

In any event, there are plenty of dual and quad opamps available that are in
many regards better than the venerable LM324. Which one is best depends on
what you're trying to do, as well as on what power supply voltages and
currents you have available. Tell us more!

 

[N. Thornton]

"V8TR4" <_v8tr4_@don't_bother_me_with_spam.com> wrote in message news:<DuVib.13470$YO5.6092568@news3.news.adelphia.net>...

Hi,

I made a small pic based project a while back that takes the outputs from
some off the shelf pressure sensors with a linear 0-10vdc output signal. The
output of the sensor goes thru a voltage devider into a MAX186 ADC, and then
into my PIC.

I have had some oscillation problems. The signals will oscillate up and down
and for some reason the signal coming in is not interpeted as being the
linear signal I expected. I added some caps around the board and the problem
became worse.

I looked into it and I see that I could use some op-amps to clean the signal
up. The MAX186 is a 0-4.096 ADC. I wanted to use the LM324, but I read that
it will give me problems due to the operating voltage. So it looks like the
741 might be a better choice. I am using all 8 channels of the ADC so eight
741's will take up a lot of board space.

Anyone have some hints and tips for what I am trying to do? What is the
cause of these oscillations and non linear numbers?

thanks in advance,
Oliver

Hi. Without more details I can only offer some general tips.

Look at your current paths, and keep the signal ground current
separate frmo the power ground current, as much as poss.

To kill an oscillating circuit one wants R rather than L or C.

You've already done the caps on the chip bit: use ceramic, as wound
caps do have L.

Finally if for whatever reason you want to use a 324, a diode or 2
could be used to shift the V range.... depends, but its possibly an
option. Really though you want to prevent the oscillation, as
filtering an oscillating detector wont give you accuracy at all.


Regards, NT

 

[V8TR4]

Hi,

Thanks for all of your responces.
I am fairly certain the oscillating is due to the capacitors on my board
that were added as an afterthought. The signal from the sensor is supposed
to be a clean linear signal. These transducer heads are used thruout
industry and most people just take the 0-10v and feed into a ADC on a
computer card. It is a selfcontained unit ready to go as they would say.

The reason why I stated I wanted to "clean up" the signals coming in was
because I thought that maybe there was some sort of an impedance problem. To
be perfectly honest, I'm not that up to speed in analog circuits. I had
heard that impedance was an issue, so figured that might be a good place to
look. I have a signal that somewhat follows the numbers given, but all of a
sudden at about 3volts the signal jumps to 2.5. pretty much out of nowhere.
What I did was I used a calibrated pressure gauge. Then ran it against the
transducer voltage and my instrument readout. I adjusted teh pressure
incrementally and recored the various readings to make comparisons.

I'm preparing to make a new board that holds the ADC and whatever opamp(s) I
decide to use. I also will provide seperate power sources for the adc as
well as the other circuits to try and minimize the problems that could occur
from that. I had everything running from a common 5vdc before. I have 24vdc,
12vdc and 5vdc available for powering my circuits. I can add a neg voltage
souce if I have to, but at this point would rather not.

How does one seperate the signal ground from the power ground? Right now I
have the grounds tied together and treat them as one. Even if I seperated
don't they eventually get back together again thru the ADC?

thanks,
Oliver


"Walter Harley" <walterh@cafewalterNOSPAM.com> wrote in message
news:vooh6of66ctu3f@corp.supernews.com...

"V8TR4" <_v8tr4_@don't_bother_me_with_spam.com> wrote in message
newsuVib.13470$YO5.6092568@news3.news.adelphia.net...
I have had some oscillation problems. The signals will oscillate up and
down
and for some reason the signal coming in is not interpeted as being the
linear signal I expected. I added some caps around the board and the
problem
became worse.

It is possible that when the voltage from the sensors increases, it
increases current draw somewhere in your system, which causes the supply
voltage (to the whole thing, or just to the sensors) to sag, which causes
the voltage output to drop proportionally, etc....

What frequency of oscillation are you talking about?


I looked into it and I see that I could use some op-amps to clean the
signal
up. The MAX186 is a 0-4.096 ADC. I wanted to use the LM324, but I read
that
it will give me problems due to the operating voltage. So it looks like
the
741 might be a better choice. I am using all 8 channels of the ADC so
eight
741's will take up a lot of board space.

I'm not at all sure what you mean by "use some op-amps to clean the signal
up." Can you be more specific?

In any event, there are plenty of dual and quad opamps available that are
in
many regards better than the venerable LM324. Which one is best depends
on
what you're trying to do, as well as on what power supply voltages and
currents you have available. Tell us more!

 

[Walter Harley]

"V8TR4" <_v8tr4_@don't_bother_me_with_spam.com> wrote in message
news:vBtjb.13872$YO5.6378938@news3.news.adelphia.net...

I am fairly certain the oscillating is due to the capacitors on my board
that were added as an afterthought. The signal from the sensor is supposed
to be a clean linear signal. These transducer heads are used thruout
industry and most people just take the 0-10v and feed into a ADC on a
computer card. It is a selfcontained unit ready to go as they would say.

The reason why I stated I wanted to "clean up" the signals coming in was
because I thought that maybe there was some sort of an impedance problem.
To
be perfectly honest, I'm not that up to speed in analog circuits. I had
heard that impedance was an issue, so figured that might be a good place
to
look. [...]

Hi, Oliver. It's going to be hard for anyone to help much unless you can
get a lot more specific. Saying that "impedance is an issue" is about like
saying "torque is an issue" when fixing a car

Let's start with a couple questions:

- Where, exactly, are these capacitors you've mentioned? And what are
their values, and what kind of capacitor are they?
- If most people are able to get this working with the transducer going
straight into an ADC, and you're not, what do you think is different about
their setups?
- What approximate frequency is the oscillation you're talking about
(closer to 1Hz or 1MHz)?
- What are the current and voltage requirements of the transducer?
- What is the output impedance of the transducer (or, what are its load
requirements)?
- You said you were using a voltage divider, presumably to scale the 0-10v
output of the transducer down to 0-4.096v for the MAX186 ADC; what value
resistors are you using?

Also, you asked about separating grounds. The quick answer here is yes, the
grounds generally connect somewhere; what you are trying to do is make sure
that current and noise from one circuit doesn't share the same ground path
as that from another. To understand this, you have to give up the idea that
ground everywhere is the same, and that a wire has zero resistance. Every
wire, every PC trace, every connector has some finite resistance and some
finite inductance (which is like resistance, at noise frequencies). So when
current flows through it, there's a voltage across it. Take your schematic,
and draw a resistor wherever there's a ground wire. Calculate the current
flow through these resistors. Now you can see what your circuit "really"
looks like, and you can start thinking about which of these resistors are
causing you a problem.

-walter