Shop OBEX P1 Docs P2 Docs Learn Events
interference? problem — Parallax Forums

interference? problem

ArchiverArchiver Posts: 46,084
edited 2003-10-22 01:31 in General Discussion
I am building a controller for my thermal solar panels.
The controller measures two temperatures and activates pumps depending on the
differential it finds.
I am using two LM35s and on the breadboard things work well, but one of the
sensors needs to be on the roof, at the other end of a 50- to 100 foot cable and
if I hook the sensor to the other end of that long line the readings become
quite erratic. Not totally off, they bounce +- 10 degrees around the real value
but obviously there is something not right.

What is the problem?

Do I need shielded cable?

TIA Uwe

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2003-10-18 01:03
    In a message dated 10/17/2003 4:43:28 PM Pacific Daylight Time,
    othello159@h... writes:
    I am building a controller for my thermal solar panels.
    The controller measures two temperatures and activates pumps depending on the
    differential it finds.
    I am using two LM35s and on the breadboard things work well, but one of the
    sensors needs to be on the roof, at the other end of a 50- to 100 foot cable
    and if I hook the sensor to the other end of that long line the readings become
    quite erratic. Not totally off, they bounce +- 10 degrees around the real
    value but obviously there is something not right.

    What is the problem?

    Do I need shielded cable?
    ____________________________________________

    When porting signals over long distances, one possible solution is to port
    current, not a voltage signal. One common range is 4 to 20 mA, where 4 mA is
    equivalent to the minimum and 20 mA the maximum.

    At your control center and the source, an op amp circuit is used to convert
    the voltage signal to current that is ported over the long distance, and
    another op amp to convert the current back into a voltage.

    Ken


    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-18 01:07
    In a message dated 10/17/2003 7:43:20 PM Eastern Standard Time,
    othello159@h... writes:


    > What is the problem?
    >
    > Do I need shielded cable

    Put a 1.2K resistor in series with the sensing lead, and put a .1uf bypass
    right at the pin connected to the sensing lead. This should stabilize your
    signal from the LM.

    Sid Weaver
    W4EKQ
    Port Richey, FL


    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-18 01:50
    Hi,

    I have an LM34 on about 75' of cable, with no problems. I used outdoor phone
    cable, unshielded. If I remember correctly, I soldered a resistor on to the
    LM34 right onto the lead before placing it in the brass tube that it lives
    in. I forget the value and placement, look on Tracy Allens's site for
    details, I did whatever it is that Tracy reccomends. It has been keeping the
    hot tub temp. perfect ever since.

    Jonathan

    www.madlabs.info


    Original Message
    From: <smartdim@a...>
    To: <basicstamps@yahoogroups.com>
    Sent: Friday, October 17, 2003 5:03 PM
    Subject: Re: [noparse][[/noparse]basicstamps] interference? problem


    > In a message dated 10/17/2003 4:43:28 PM Pacific Daylight Time,
    > othello159@h... writes:
    > I am building a controller for my thermal solar panels.
    > The controller measures two temperatures and activates pumps depending on
    the
    > differential it finds.
    > I am using two LM35s and on the breadboard things work well, but one of
    the
    > sensors needs to be on the roof, at the other end of a 50- to 100 foot
    cable
    > and if I hook the sensor to the other end of that long line the readings
    become
    > quite erratic. Not totally off, they bounce +- 10 degrees around the real
    > value but obviously there is something not right.
    >
    > What is the problem?
    >
    > Do I need shielded cable?
    > ____________________________________________
    >
    > When porting signals over long distances, one possible solution is to port
    > current, not a voltage signal. One common range is 4 to 20 mA, where 4 mA
    is
    > equivalent to the minimum and 20 mA the maximum.
    >
    > At your control center and the source, an op amp circuit is used to
    convert
    > the voltage signal to current that is ported over the long distance, and
    > another op amp to convert the current back into a voltage.
    >
    > Ken
    >
    >
    > [noparse][[/noparse]Non-text portions of this message have been removed]
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    >
    > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-18 02:53
    http://www.emesys.com/tf_diy.htm

    Shows how to clean up the bouncy signal. He even sells a really nice parts
    kit to make them.

    >
    Original Message
    > From: othello159 [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=2OeMUzanJhXTldJkgXvywojVt7ECAMDqbbNCOUww-mPTKHDmQ1LmLjYdtNfvVUCWjNK_CgWV-6S6pRef]othello159@h...[/url
    > Sent: October 17, 2003 4:41 PM
    > To: basicstamps@yahoogroups.com
    > Subject: [noparse][[/noparse]basicstamps] interference? problem
    >
    >
    > I am building a controller for my thermal solar panels.
    > The controller measures two temperatures and activates pumps
    > depending on the differential it finds.
    > I am using two LM35s and on the breadboard things work well, but
    > one of the sensors needs to be on the roof, at the other end of a
    > 50- to 100 foot cable and if I hook the sensor to the other end
    > of that long line the readings become quite erratic. Not totally
    > off, they bounce +- 10 degrees around the real value but
    > obviously there is something not right.
    >
    > What is the problem?
    >
    > Do I need shielded cable?
    >
    > TIA Uwe
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the
    > Subject and Body of the message will be ignored.
    >
    >
    > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-21 01:11
    --- In basicstamps@yahoogroups.com, Newzed@a... wrote:
    > In a message dated 10/17/2003 7:43:20 PM Eastern Standard Time,
    > othello159@h... writes:
    >
    >
    > > What is the problem?
    > >
    > > Do I need shielded cable
    >
    > Put a 1.2K resistor in series with the sensing lead, and put a .1uf bypass
    > right at the pin connected to the sensing lead. This should stabilize your
    > signal from the LM.
    >
    > Sid Weaver
    > W4EKQ
    > Port Richey, FL


    Thanks for all the suggestions, I will try some of them out in the next couple
    of days.

    But why should this work ( I always want to know why). I can see how the cap
    would drain unwanted AC to ground, but why the resistor. Do you know the theory
    behind it?

    regards

    Uwe Langmesser
    KB1JOW
    Maine, USA
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-21 01:19
    In a message dated 10/20/2003 8:12:51 PM Eastern Standard Time,
    othello159@h... writes:


    > But why should this work ( I always want to know why). I can see how the
    > cap would drain unwanted AC to ground, but why the resistor. Do you know the
    > theory behind it?
    >

    The cap gets rid of noise and other ibnterference. I don't know what the
    resistor does - I just know that it works. I build a 40 LM34 system for a
    customer with runs from 5 feet to 36 feet and the resistor/cap combinations
    solved
    all my instability problems.

    Sid


    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-21 04:09
    --- In basicstamps@yahoogroups.com, Newzed@a... wrote:
    > In a message dated 10/20/2003 8:12:51 PM Eastern Standard Time,
    > othello159@h... writes:
    >
    >
    > > But why should this work ( I always want to know why). I can see how the
    > > cap would drain unwanted AC to ground, but why the resistor. Do you know the
    > > theory behind it?
    > >
    >
    > The cap gets rid of noise and other ibnterference. I don't know what the
    > resistor does - I just know that it works. I build a 40 LM34 system for a
    > customer with runs from 5 feet to 36 feet and the resistor/cap combinations
    solved
    > all my instability problems.
    >
    > Sid
    >

    That is a lot of sensors....

    Have you noticed any of this with your sensors:
    It takes my sensors sometimes a while to "settle". Debug shows me they switch
    back and forth between two temp reading. That leads to my pumps switching on and
    off for a number of times before they settle in the on or off state. It is not a
    huge problem but I wonder how to stabilize this.
    Putting a pause command in the software to read the temp only every 5 sec.
    helped a bit but didn't stabilize the system entirely.

    Uwe
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-21 06:09
    >--- In basicstamps@yahoogroups.com, Newzed@a... wrote:
    >> In a message dated 10/20/2003 8:12:51 PM Eastern Standard Time,
    >> othello159@h... writes:
    >>
    >>
    >> > But why should this work ( I always want to know why). I can see how the
    >> > cap would drain unwanted AC to ground, but why the resistor. Do
    >>you know the
    >> > theory behind it?
    >> >
    >>
    >> The cap gets rid of noise and other ibnterference. I don't know what the
    >> resistor does - I just know that it works. I build a 40 LM34 system for a
    >> customer with runs from 5 feet to 36 feet and the resistor/cap
    >>combinations solved
    >> all my instability problems.
    >>
    >> Sid
    >>
    >
    >That is a lot of sensors....
    >
    >Have you noticed any of this with your sensors:
    >It takes my sensors sometimes a while to "settle". Debug shows me
    >they switch back and forth between two temp reading. That leads to
    >my pumps switching on and off for a number of times before they
    >settle in the on or off state. It is not a huge problem but I wonder
    >how to stabilize this.
    >Putting a pause command in the software to read the temp only every
    >5 sec. helped a bit but didn't stabilize the system entirely.
    >
    >Uwe

    Hi Uwe,

    The LM34 is a micro-power circuit that uses feedback internally to
    stabilize its voltage output. It is a common problem with of this
    type, that the output will be unstable when it has to drive a cable,
    or a capacitor of a certain size. One of the characteristics of this
    instability is that it will seem to flop back and forth between two
    temperature readings, unpredictably. If you look at the output
    signal on an oscilloscope, you can often catch a wild oscillation at
    a frequency of several thousand hertz, and 100 mV or more amplitude.
    At other times that oscillation will not be there. It may go away
    when you touch the output with the 'scope probe or with your finger,
    or when you power the circuit up one time but not another.

    The combination of resistor and capacitor at the output tames the
    oscillation. Intuitively, it works something like this. Say a
    fluctuation occurs in the temperature, which internally is a small
    voltage change that is amplified to the output. But there is a bit
    of a delay and ramp up of the response at the output, and that slow
    response at the output becomes more acute when a capacitor or a cable
    is attached. Feedback from the output to the input senses that the
    output is not changing fast enough and and tries to turn the output
    on harder, again with a delay or "lag" to the output. The result is
    that the voltage at the output overshoots, and under conditions that
    are all too easy to find, the overshoot becomes sustained
    oscillation. The resistor you add at the output generates a voltage
    proportional to the current flowing out into the load capacitor and
    that additional "leading" voltage stabilizes the feedback loop.

    You can read about this in the LM34 data sheet, and I also have
    information on my web pages at
    http://www.emesys.com/OL2heat.htm#TF_LM34

    -- regards,
    Tracy Allen.
    electronically monitored ecosystems
    mailto:tracy@e...
    http://www.emesystems.com
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-22 01:11
    --- In basicstamps@yahoogroups.com, Tracy Allen <tracy@e...> wrote:
    > >--- In basicstamps@yahoogroups.com, Newzed@a... wrote:
    > >> In a message dated 10/20/2003 8:12:51 PM Eastern Standard Time,
    > >> othello159@h... writes:
    > >>
    > >>
    > >> > But why should this work ( I always want to know why). I can see how the
    > >> > cap would drain unwanted AC to ground, but why the resistor. Do
    > >>you know the
    > >> > theory behind it?
    > >> >
    > >>
    > >> The cap gets rid of noise and other ibnterference. I don't know what the
    > >> resistor does - I just know that it works. I build a 40 LM34 system for a
    > >> customer with runs from 5 feet to 36 feet and the resistor/cap
    > >>combinations solved
    > >> all my instability problems.
    > >>
    > >> Sid
    > >>
    > >
    > >That is a lot of sensors....
    > >
    > >Have you noticed any of this with your sensors:
    > >It takes my sensors sometimes a while to "settle". Debug shows me
    > >they switch back and forth between two temp reading. That leads to
    > >my pumps switching on and off for a number of times before they
    > >settle in the on or off state. It is not a huge problem but I wonder
    > >how to stabilize this.
    > >Putting a pause command in the software to read the temp only every
    > >5 sec. helped a bit but didn't stabilize the system entirely.
    > >
    > >Uwe
    >
    > Hi Uwe,
    >
    > The LM34 is a micro-power circuit that uses feedback internally to
    > stabilize its voltage output. It is a common problem with of this
    > type, that the output will be unstable when it has to drive a cable,
    > or a capacitor of a certain size. One of the characteristics of this
    > instability is that it will seem to flop back and forth between two
    > temperature readings, unpredictably. If you look at the output
    > signal on an oscilloscope, you can often catch a wild oscillation at
    > a frequency of several thousand hertz, and 100 mV or more amplitude.
    > At other times that oscillation will not be there. It may go away
    > when you touch the output with the 'scope probe or with your finger,
    > or when you power the circuit up one time but not another.
    >
    > The combination of resistor and capacitor at the output tames the
    > oscillation. Intuitively, it works something like this. Say a
    > fluctuation occurs in the temperature, which internally is a small
    > voltage change that is amplified to the output. But there is a bit
    > of a delay and ramp up of the response at the output, and that slow
    > response at the output becomes more acute when a capacitor or a cable
    > is attached. Feedback from the output to the input senses that the
    > output is not changing fast enough and and tries to turn the output
    > on harder, again with a delay or "lag" to the output. The result is
    > that the voltage at the output overshoots, and under conditions that
    > are all too easy to find, the overshoot becomes sustained
    > oscillation. The resistor you add at the output generates a voltage
    > proportional to the current flowing out into the load capacitor and
    > that additional "leading" voltage stabilizes the feedback loop.
    >
    > You can read about this in the LM34 data sheet, and I also have
    > information on my web pages at
    > http://www.emesys.com/OL2heat.htm#TF_LM34
    >
    > -- regards,
    > Tracy Allen.
    > electronically monitored ecosystems
    > mailto:tracy@e...
    > http://www.emesystems.com


    Thank you very much for explaining this.
    Before I received your mail I had placed the 0.1 mF capacitor and the resistor
    in the circuit as was suggested. That had stopped the rather erratic and wide
    range change of numbers, but I was still getting the (much more tolerable)
    "threshhold jitter " when temperature changes caused the LM 35 to change from
    one number to the other.
    On a hunch I placed a larger electrolytic capacitor, 10 mF, parallel to the 0.1
    and the set up has been working stable as a rock ever since.

    Sometimes bigger is better...

    Uwe
  • ArchiverArchiver Posts: 46,084
    edited 2003-10-22 01:31
    >Before I received your mail I had placed the 0.1 mF capacitor and
    >the resistor in the circuit as was suggested. That had stopped the
    >rather erratic and wide range change of numbers, but I was still
    >getting the (much more tolerable) "threshhold jitter " when
    >temperature changes caused the LM 35 to change from one number to
    >the other.
    >On a hunch I placed a larger electrolytic capacitor, 10 mF,
    >parallel to the 0.1 and the set up has been working stable as a rock
    >ever since.
    >
    >Sometimes bigger is better...
    >
    >Uwe


    Hysteresis in software can also help stabilize the switching:

    IF temperature > threshold THEN
    GOSUB gizmo_on
    ELSEIF temperature < threshold THEN
    GOSUB gizmo_off
    ENDIF


    That provides a two unit dead band that suppresses jitter.


    ;-<
    ;---ON---
    ___OFF_______|___>___|
    T-1 T T+1
    -- Tracy
Sign In or Register to comment.