• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 47 subscribers
  • Views 709 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TI LM139AD Hysteresis

Anthony Campos
Intellectual 400 points
Other Parts Discussed in Thread: LM139

Using the LM139AD and simulating using TINA with Hysteresis, however the results are not as expected.  Using a non inverting configuration with R2 = 10M, and R1 = 290K. My Vref is 12V, so this brings my top trip point to around 15V.  However, I want the bottom trip point to be 9V, but for some reason it's around the same amount as the rising trip point.  My output is being pulled up to 3.3V, however my Vcc is 28V so my input can be 0 to 28V.  Is there a way to set the lower trip point? I want 3V of hysteresis above and below 12V.

  • 0
    Guru 117855 points

    Key here is likely that your output is pulled up to only 3V3.  Using comparator hysteresis - you rely on the difference in levels between comparator "on" and "off."  As your comparator output may "swing" only ~ 3V - it will not be possible to create the 6V swing you seek - employing "just" Rfeedback (R2) and Rinput (R1).  You would have to implement additional circuit paths and components to achieve your wide ranging trip points.

    One method may be to allow the comparator's output to swing the full 28V - and limit just your applied output to 3V3.  (a simple voltage divider downstream from the comparator's output should easily achieve)

  • 0 in reply to cb1_mobile
    Intellectual 400 points

    I'm thinking maybe I don't need a 6V swing.  More likly a 3V swing around 12V.  Do you think this would be possible? and how does one achieve the lower trip point? Does it require more than just R2 and R1?

  • 0 in reply to Anthony Campos
    Guru 117855 points

    We've used such analog comparators for past 20+ years - have never required so high a hysteresis swing.  Usual purpose of hysteresis is to prevent oscillation on/around the switching point - usually 100-300mV of hysteresis is more than adequate. 

    May well be that you can better achieve your objective by using a 2nd section of this 4 section comparator - having one trip @ 13.5V and second @ 10.5V.  Use external logic to detect the appropriate condition.  (note - comparator ap notes show how one may achieve "and, or, and invert" logic - using another section of same comparator.)

    The more you allow the comparator output to "swing" - the greater the potential range of your hysteresis.  As the comparator is open collector - you can pull-up the output to just about anywhere (positive) so long as you don't exceed Vcc.  As stated - voltage divider can limit output high to 3V3 for your target circuit - even though the comparator output may yield a logic high of 12-24V. 

    The ratio between feedback & input resistors along with the difference in comparator output voltage (Voh vs Vol) will determine hysteresis.  Again - suspect the broad range you seek is best achieved via 2nd section of your quad comparator.  You may wish to implement a "window comparator" which will "report" when your input signal is w/in the range "greater than Vlo-set and less than Vhi-set."  Two comparator sections can so implement...

     

  • 0 in reply to cb1_mobile
    Intellectual 400 points

    Yes, I'm used to seeing much lower hysteresis myself.  I saw 3V hysteresis on this other vendor's discrete interface IC.  So if I ultimately select 100-300mV hysteresis I will still only have one trip point correct? Is this because I have a single supply comparator?

  • 0 in reply to Anthony Campos
    Guru 117855 points

    Continue to believe my earlier post fully guides - very doubtful that 3V is "hysteresis" - sounds much like IC's switching "level" - which is not hysteresis.  Again - hysteresis usually comes into play when the input signal changes relatively slowly - and may "hang" at/around the switching threshold - causing the output to "chatter."

  • 0 in reply to cb1_mobile
    TI__Guru 68541 points

    Anthony,

    One of the thing you have to be careful about while using comparators is not to exceed the input common-mode voltage range as you set the trip point; doing so will virtually shut-off one side of the input differential pair resulting in very high Vos (perhaps >1V) that changes dramatically depending on the input voltage at the other input terminal.  This could be perceived as very large input voltage hysteresis but it has actually nothing to do with a true hysteresis (commonly used in de-glitching comparators) but it's rather result of a non-working input differential pair.  LM139 valid input common-mode voltage ranges from a negative rail ONLY up to Vcc-2V so you may NOT set a trip point closer than 2V below a positive rail - see PDS table below.

     

     

    If this is not what causes your problem, instead of describing the circuit please provide us with the circuit diagram showing supply voltages, component values and input voltages.

  • 0 in reply to Anthony Campos
    TI__Guru**** 166906 points

    Hello Anthony,

    A 6V hysteresis swing is still possible even with a 3.3V pull-up voltage.

    The amount of hysteresis is (Voh- Vol) * Rin / Rf
    Swing is 6V, Voh is 3.3V, and Vol is zero. So Rin/Rf is 1.82

    Rf could be 10k and Rin could be 18.2k
    Now, the reference voltage on the inverting input must be calculated.
    Vref = [ (VT+)(Rf) - (Rin)(VOL) ] / (Rin+Rf)
    Vref = [ 15V*10k - 18.2k*0V ] / 28.2k
    Vref = 150kV / 28.2k = 5.32V

    A resistor divider can lower your 12V to 5.32V for the inverting input.
    The final circuit is non-inverting and will switch output states at 9V and 15V

    Regards,
    Ron Michallick

     

     

     

  • 0 in reply to Ron Michallick
    TI__Guru**** 166906 points

    Anthony,

    There is one precaution. The input voltage will have an effect on VOH.
    The pull up resistor should be at least 100 times less than the sum of the input and feedback resistors.
    In the example, that would be less than 282 ohms. So it would be good to increase the Rin and Rf resistors.
    To prevent IIB from causing a threshold shift, I suggest that the total resistances on both inputs be matched.

    Ron