voting logic circuit design ideas?

Discussion in 'The Projects Forum' started by Mike the Tinkerer, Dec 20, 2007.

  1. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Hello,

    I am new to the site and would greatly appreciate some help on a project.
    I understand the basics of Electronics and circuits (NAVY Training years ago), but an accomplished tech or programmer I am not. Please keep this in mind, and assume my skills as that of a 3 year old :).

    I would like to make a circuit that takes 3 or 4 sensor inputs, compares them, and "votes out" the sensor that is not responding within specs + or - 20% of comparative reading from other sensors. The final good output would need to be averaged out between the "good" sensors,and sent to a LED display.

    The sensor output is between 10mv and 100mv.

    Thank you all for your help on this.

    Cheers :D

    Mike
     
  2. recca02

    Senior Member

    Apr 2, 2007
    1,211
    0
    i think this would be easier to accomplish with programming...and believe me i m worse than u when it comes to that;).

    some of the members here might need some clarifications like:
    again does the comparison take place between the average of other three signals and the fourth one? the +-20% will change for each signal else wont it?
    how many must be voted out? can they be more than one?
     
  3. Out To Lunch

    New Member

    Dec 7, 2007
    3
    0
    are you expecting the sensors to begin failing after a while? is the environment they are going in volatile or are they just cheap/fragile sensors? If you expect them to fail, what is your assurance that the one that looks like it is failing really is the one that is failing and not the other way around?

    if you expect aberrant reading with perfectly good sensors, i'm not sure why you would want to ignore the data. if the sensor is good and it is reading differently than the others, then it may be indicative of something that warrants attention.
     
  4. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Let's assume we are dealing with 3 sensors.

    The circuit monitors the voltage outputs of all 3 sensors. Let's say after 1 hour, 2 of the sensors read 20mv and 1 reads 31mv, the sensor reading 31mv would be "voted out" and therefore eliminated from further comparison between the sensors. The circuit now averages the 2 remaining sensors and displays the results on a LCD (Need help with a lcd driver too :rolleyes:). The "voted out" sensor stays out of further calculations until the unit is turned off. The circuit cannot have less than 2 sensors active all all times, so only 1 sensor would ever be voted out until reset.

    I hope I'm making sense.

    Thanks guys.

    Cheers :D

    Mike
     
  5. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Yes the environment is very volatile. The sensors are subject to failure fairly frequently. The voting logic, along with a separate readout of the MV output of each sensor is the only way that I know to monitor and control the project.

    This is for an environmental control unit that I am building. The sensors are in a o2 rich environment with varying pressures up to 300 psi and humidity up to 100%.

    Thanks again,

    Cheers :D

    Mike
     
  6. cumesoftware

    Senior Member

    Apr 27, 2007
    1,330
    10
    Definitely a computer program would be a better idea. It is an expandable solution since it may provide flexibility to the user (i.e. the option of saving logfiles).
     
  7. Salgat

    Active Member

    Dec 23, 2006
    215
    1
    A micro controller has the ability to convert voltages into digital values. You can process these values as you please to control the circuit.
     
  8. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Is this something that I could build in a 3" x 3" space? Controller has to be really small

    Thank you :D
     
  9. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0

    Any specific one come to mind? If you have a part number or Model, I can start looking it up and trying to figure it out.

    Thanks again

    Cheers :D

    Mike
     
  10. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    How much money expressed in US dollars are you comfortable spending?

    hgmjr
     
  11. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    I have a $500 budget for the sensor project, the pressure sensor interaction, calibration routine, and ad/da contol of a scelonoid.

    This does not include hard parts, just the programming, and R&D.

    This is a personal project, where I need precise control of an environment to test materials. Hence the fact that it needs to operate in o2 environments up to 100%, high pressure up to 300psi, and humidity up to 100%. It also needs to be small, and with low battery draw (except for the LCD).

    What do you have in mind?

    Cheers

    Mike
     
  12. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Several questions:

    When do you need to be up and operational?

    Are you interested in learning the needed electronics?

    Are you interested in learning to program in C-language?

    These are not tasks for the "faint of heart" but it is safe to say that there is a significant payoff for the time and effort invested.

    hgmjr
     
  13. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Realistically, I would like to have this operational within 6 months or so.

    That would be fine, if that's what I need to do to accomplish this goal.

    I am not familiar with programming in C or C++, I do have some limited experience in Basic. I was looking at basic stamp programming, with a/d converters for the sensors. Am I on the right track?

    Yes I can see that, but I am always up for a challenge. I wish I worked for a large corporation with on staff programmers, but unfortunately I am an automotice repair tech, and although I have "some" experience with electronics, it is nowhere near the wealth of knowledge that you guys have.

    Oh, and I can't tell my wife where the $500 is going :rolleyes:

    Thanks a bunch. :D

    Cheers

    Mike
     
  14. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    You mentioned that your operating environment is 100% humidity which if I am not mistaken is the equivalent of being submersible. What is the temperature range that must be operated within?

    Isn't the O2 environment explosive?

    hgmjr
     
  15. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0
    Yes, it is equivalent to being submersed, but I was planning on sealing the module against moisture if possible. The sensors would be exposed to moisture although not submersed, more like condensation, but that's still considered 100% humidity.

    It should be stable between -10f and +120f

    Yes it can be, that is why I would like to seal the controller against any contamination with pure o2. The sensors are inert, therefore not explosive.

    Cheers :D

    Mike
     
  16. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    You mentioned the need for high precision. Can you quantify what level of resolution you need to achieve using an example of a typical measurement that you would be making?

    hgmjr
     
  17. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0

    The precision on the programming end is really limited to the relative % of deviation from one sensor to the other. Let's assume 30%. The sensors themselves have a +-5% deviation in reading, therefore if a sensor decides to shoot itself in the head and deviate from the others in more than a 30% value, it would be voted out and no longer included in the average reading from the other 2.

    As long as we know the +- drift of the sensors, all the circuit has to do is monitor the MV output of the sensors and compare them. I believe the sensor outputs are from 1 to 100mv.

    Thanks

    Cheers :D

    Mike
     
  18. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    I am not that familiar with the BASIC Stamp's features as far as A-to-D conversion is concerned. I am more familair with the ATMEL AVR microcontroller series which I know has a 10-bit A-to-D converter together with the ability to multiplex several channels into the A-to-D. 10-bits allows one to discern 1 part in 1024 though it is more realistic to depend only on 1 part in 512. That should be adequate for your application.

    I am pretty certain that the MICROCHIP's PIC series of microcontrollers is similarly endowed.

    Based on the maximum output level of your sensor, you are going to need an opamp to boost the signal level by 40 to 50 times.

    hgmjr
     
  19. Mike the Tinkerer

    Thread Starter Member

    Dec 20, 2007
    14
    0

    Thank you for the input, I will look into the Basic stamp and the AVR, Although I believe that the AVR will be overkill for my project. What do you recommend as an efficient Opamp?

    Thanks

    Cheers :D

    Mike
     
  20. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    There are many opamps from which to chose. Can you tell us more about the characteristics of the sensor being used? You told us the output range is 1 to 100 mv over the range that you plan to use it. Do you have a datasheet or perhaps a part number that can be used to look up the specs for the sensor.

    I am at least somewhat concerned about the suitability of using batteries as the power source for your project. The temperature range is my first concern. I don't know that batteries are going to hold up well at the temperature extremes you have mentioned. The other concern is the high pressure of 300 psi. I don't have any direct experience with the use of batteries in high pressure applications.

    Maybe one of our other members can speak to the use of batteries in high pressure applications. There may be a battery technology that is optimized for such applications.

    hgmjr
     
Loading...