Connect with us

Hysteresis needed in 741 Op-Amp circuit

Discussion in 'Electronic Basics' started by CF, Jun 10, 2004.

Scroll to continue with content
  1. CF

    CF Guest

    Well, I did end up putting a voltage regulator in there, though I went for
    a 7812 1 amp unit because of the fan that I want to power too. It seems to
    make the relay very jumpy when it's switching off. Something about the
    addition of the 7812 seems to sensitize the circuit more. I can tune it out
    a bit with more feedback, but this widens my hysteresis zone beyond 1
    degree C, which I don't really want. So I'll ditch it and go for resistors
    on the relay coil and fan circuits, which is neat enough anyway. By my
    measurment, the 7812 draws 3.7mA just to power itself, so I'm not overly in
    love with it (speaking in perfectionistic terms, for the sake of it).

    Measurements are as follows:


    AC from transformer: 14.75V

    DC with no load from fan: 17.9V

    DC with fan drawing: 16.0V

    Current of entire circuit, quiescent: 2.3mA AC, 1.9mA DC

    Current of entire circuit, relay coil on: 66mA AC, 57mA DC

    Current of entire circuit, relay & fan: 236mA AC, 220mA DC

    Casualties: One 0.5A multimeter fuse, 6km round-trip


    Voltage at all times: 11.9V DC

    Current of entire circuit, quiescent: 5.0mA DC

    Current of circuit minus 7812, quiescent: 1.3mA DC

    Current of entire circuit, relay coil on: 9.5mA DC (strange!)

    Current of entire circuit, relay & fan: 171mA DC

    Casualties: Another 0.5A fuse, another 6km round-trip tomorrow.

    Lesson: Buy two fuses.

  2. John Fields

    John Fields Guest

    Something I should have pointed out, (but didn't, because I didn't
    think about it) was that the LM311 (or your 741, for that matter)
    _can't_ run on unfiltered DC, since it'll get unhappy once the output
    from the rectifier falls below its minimum allowed positive supply
    voltage, duhh...
    Milli_watts_? Surely you meant milli_amps_ of secondary current,
    which would make the transformer rated for something like 12V * 150mA
    = 1.8VA. Plus, there's the cost of running the capacitive input
    filter, which can run up to about 60% of the load, which means you
    should be using about a 3VA transformer.

    With a 2200µF cap and a 160mA load (the relay _plus_ the fan) you can
    expect a ripple voltage across the cap of about +/-0.35V on top of the
    DC out of the filter.

    A good idea, when you're asking questions, is to include as much
    information as possible about what your goal is, what you've done, so
    far, to get there, and what you've measured in order that as little
    ground as possible has to be covered more than once. That saves us
    _all_ time, effort, and heartburn.
  3. John Fields

    John Fields Guest

    Since the side of the 634k resistor connected to the thermistor will
    be sitting at about 6V and the other side of it will be connected to
    essentially the unloaded peak voltage out of the bridge when the relay
    is off, (19V, I think you said) the load on the cap will be about 20µA
    so, at 100Hz, the ripple voltage across it will be about 21mV, which
    shouldn't interfere seriously with your hysteresis. Also, since the
    impedance of the source charging the cap is negligible in this
    instance, you could use a vanilla aluminum electrolytic. Any 10µF cap
    with a voltage rating greater than the unloaded output voltage from
    the bridge would work. 25V would probably be OK, but for a few
    pennies more I'd use 35V just to be safe.
  4. John Fields

    John Fields Guest

    Buy a box of 'em and save yourself a lot of trips!^)

    Offhand, I'd say that if you measured 9.5ma with the relay on, you
    either measured the current incorrectly or you didn't install the 7812
    properly. Moreover, you should check to see whether the fan can run
    properly with the full-wave rectified AC you've already got (instead
    of on filtered DC) since if it will that'll save you an electrolytic
    cap and take an additional load off of the transformer, since it won't
    have to charge that cap if it isn't there.

    Assuming that your fan can run on rectified, unsmoothed DC (I just
    checked three different brushless DC fans and they all ran just fine,
    so my guess is that it can) and that the fan will run when the relay's
    on, your circuit (my LM311 version, I don't even want to mess around
    with the 741 stuff) should look like this:

    12VAC>--------------------+ |
    | |
    | | |
    | |
    | | |
    | IN +------+
    +-------------------[78L12] | |
    | OUT | |
    | | [R3] [R4]
    | | | |
    | | | |
    | +--------+--------+ +-------+ |
    | | | | |K | O---> |
    | [R1] [10.7K] | [1N4001] [COIL]- - - - -|
    | | | | | | O
    | | +-[634K]-------+-------+ |
    | | | | | |
    | | | +------+ | |
    | | +----|+ C|----------+ |
    | | | |LM311 | | [FAN]
    | [1000]<-----|----|- E|--+ | |
    | | | +------+ | |+ |
    | | | | | [10µF] |
    | [R2] [RT1] | | | |
    | | | | | | |

    If you're worried about burning up the fan or the relay, measure the
    DC current into them (separately) and then adjust R3 and R4 until you
    get their rated currents into them, then that should be it.
  5. CF

    CF Guest

    Good idea!

    Some strange thing with the multimeter I think. Cosmic rays maybe.

    Excellent John, thanks for this. I didn't realize I had all those options
    of feeding different DC here and there. For the sake of simplicity at the
    moment, I've just used the smoothed, rectified DC everywhere, but may
    fiddle with it later, depending on whether I switch transformers, etc.

    Today I've taken the plunge and wired the 311 in. It works! My circuit is
    slightly different to the above, but at least it's going. I can refine it
    later. I've got resistors in line with the relay and fan, so things are
    nice there.

    Yes, typo. The total AC in now is around 180mA with the fan running. What
    are the likely consequences of using a 150mA transformer (assuming that's
    what it is)?

    I've quickly posted my circuit as wired at the following page, less the
    resistors mentioned above that are peculiar to my situation:

    Kind regards,

  6. John Fields

    John Fields Guest

    The 1N4148 you've shown connected from the bottom of the relay coil to
    ground isn't doing anything useful and could allow the 311's output
    stage to be damaged/destroyed by the relay's turn-off transient. To
    suppress the transient, that diode should be a Zener rated at a
    voltage higher than the supply voltage, but less than the the maximum
    voltage allowed across the 311's output. 25V should be fine. The
    other (possibly more convenient) way to do it is to connect the 1N4148
    across the coil with the diode's cathode connected to the supply.
  7. CF

    CF Guest

    Ah yes, how silly of me. I was scratching my head about that one trying to
    understand how it could possibly function in the same way as a diode across
    the coil. I see now that it was the zener side of things that was doing the
    work in that configuration. Thanks for putting me straight. I've fixed both
    the diagram and real-life circuit.

    Thanks for the capacitor info too.

  8. CF

    CF Guest

    Well, I want to follow-up for the record...

    I've committed heresy and reverted back to the LM741 op amp version of the
    circuit! I found that for my application, with a remotely-mounted
    thermistor and temperature dial, perfboard construction, and wires going
    everywhere (including mains), the 311 was just too prone to stray
    oscillations. A fine comparator it is, but it's just too good for this job.

    I was not able to get less than about 1 degree celsius of hysteresis
    without the 311 going into convulsions. Even at that point, switching
    things on and off and moving things would sometimes be enough to set it
    twitching. I tried to cure this, but in the end decided it wasn't worth
    messing around with when the 741 worked so nicely.

    With the 741 and a 1 Meg feedback trimpot I can get down to about 1/3
    degree hysteresis at 25C, which is great. It's stable. It draws a good
    milliamp less, so why not use it? Horses for courses.

    This interesting diversion and happy outcome would not have been possible
    without the great input I've had here -- so many thanks to everyone.

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day