How can we help?


Pro-16 Flow Meter count increments about once every second without water running

Comments

16 comments

  • Avatar
    Istvan

    Hi Glenn,

    That 5V output is protected against overcurrent with the limit being around 35mA. If 2.2K works, I recommend using that to save power. Having a 200ft cable might result in picking up noise along the way - a pull-down to ground is a good way to make sure you filter out noise.

    >> I have also found that connecting G to Ground works
    What is "Ground"? Protective earth?

    Thank you!
    Istvan
    RainMachine Support

  • Avatar
    kujulu

    Thanks, Glenn, for posting this, as it happens intermittently on my Pro-16 with the Hydrawise HC-100 flow meter which is also a simple reed switch design. I'm "glad" there's a solution, but curious why the Pro-16 needs it but not, say, a matching Hunter irrigation controller.

    BTW when I say "intermittently" I mean the "gallons leaked" is some 60x greater than the actual gallons sent out the irrigation — yet the Pro-16 doesn't always detect false leaks.

    I can also confirm that when the Pro-16 detects false leaks, it ticks away even with just the F wire connected (i.e. 5V disconnected).

    So Istvan's comments about noise sounds plausible, though in my case the cable bundle (unshielded, and not ethernet) is only about 10 feet in length. I wouldn't have thought shielding was necessary at that short distance ....

    > I have also found that connecting G to Ground works. 

    I too am interested in what you mean. Do you mean, for example, connecting G to C?

  • Avatar
    Glenn Verner

    I used the power supply from my previous Rain Bird timer which had a three prong plug and and additional green wire that I connected to G. 

  • Avatar
    kujulu

    Thanks Glenn. 

    So to be sure I understand, either a resistor across F and G or G directly to the house’s ground works to eliminate false detection of leaks?

  • Avatar
    Glenn Verner

    Yes 

  • Avatar
    kujulu

    Thanks Glenn.

    Istvan: Does Rainmachine have a sense of which resistance/power resistor would be most ideal for avoiding these false leak detections? Unlike Glenn I'm not naturally endowed with resistors lying around, so I get to choose which one to buy!

  • Avatar
    Istvan

    Hi kujulu,

    I will try to reproduce the misbehavior and come back with a suggestion for you, that 2.2k seems like a good value, though.
    I will share my findings here, stay tuned.

    Thank you,

    Istvan
    RainMachine Support

     

  • Avatar
    Istvan

    Hi again,

    I haven't been able to reproduce the behavior you experience with just a long cable attached - I will try some more experiments with routing the wire around possible noise sources and see what I get. You can use a 2.2K, that pull-down is strong enough and still far from the current limit of the 5V output.

    Istvan,
    RainMachine Support

  • Avatar
    kujulu

    Thanks for pursuing this Istvan. My last false leak was detected on June 24 (11 days ago), but by that time my Pro-16 had detected 19,323,990.8 gallons of leaks vs an overall water consumption of 3757.6 of irrigation water consumption.

    It would be good to figure out what's causing this. Given your testing, is it possible this was resolved in production and that an RMA would help in my (and Glenn's) case?

  • Avatar
    Istvan

    Hi!

    This just hit me, I don't know how I missed it: the 60x multiplier must be because of a serious crosstalk or an occasional short between the sensor input and the common wire (which is actually one of the 24VAC 60Hz wires of the transformer). A resistor wouldn't help in the case of a short, though, so it must be the crosstalk. How consistent was this 60 multiplier, can you remember? 

    Thanks for your input!
    Istvan

  • Avatar
    kujulu

    > How consistent was this 60 multiplier, can you remember? 

    Not consistent, I'm afraid. My latest actual figures are above and the multiplier had "grown" to over 5000x.

    The runaway false leak detections occur unpredictably. In that regard, it could still be a shielding or general crosstalk issue.

    Hunter/Hydrawise does recommend using shielded cabling for its flow meters (like my HC-100) but my installer has never needed it before when using Hunter's own computers.

  • Avatar
    Istvan

    Hi,

    Yes, the shielding makes sense and it would explain why connecting the GND to protective earth helps.
    Did you add the pull-down resistor between G and F? Do you still see the random ticks happen?

    >> It would be good to figure out what's causing this. Given your testing, is it possible this was resolved in production and that an RMA would help in my (and Glenn's) case?
    We haven't made any hardware modifications on the sensor input circuitry, so i don't think so. 

    Istvan
    RainMachine Support

  • Avatar
    kujulu

    Hi Istvan,

    > Did you add the pull-down resistor between G and F? Do you still see the random ticks happen?

    Not yet, because I haven't gotten a false leak detection since resetting that counter at the time of my most recent post three days ago. I wanted to see what the frequency of detections was like before applying the resistor "fix".

    This problem is so hard to debug, as it happens apparently randomly. It used to happen more frequently until we completely clipped off the HC-100 flow meter's red lead, which isn't supposed to be used. It had previously been connected to the bundle of cables going to (but not connected to) the Rainmachine. This is also the reason I think shielding may be the issue.

    But when the false detections start, they don't stop until maybe the next program runs. That's how you get millions of gallons "leaked", as you see above.

    I'll update when I can test the resistor hack. Other than additional power draw what would be the potential side effects?

    Also, what do you think about Glenn's tying G to the house's electrical ground?

  • Avatar
    kujulu

    After three blissful (but annoying) weeks with no false leaks detected, they came back unremittingly in force today.

    I dug up a tiny (1/8W?) 20 year-old 2.4k resistor and installed it between F and G. Et voila: the runaway "leaks" instantly stopped.

    What capacity resistor do I need for this fix?

    Any idea on why this hack is even necessary to avoid false leak detection?

  • Avatar
    Istvan (Edited )

    Hi Kujulu,

    A 1/8W rated resistor is quite enough for this (I calculated a value around 0.01W).
    Do you have some bigger consumer around your RM? One that runs rarely but could produce radiated noise the RainMachine might pick up?
    I think that - unless this hack is made - both the GND and the F nets act as antennas in your setup. Having the GND connected to protective
    earth might help as you effectively shunt picked-up noise to earth, not allowing it to get further through the input circuitry (on either nets).

    Anyway, I will take my setup outdoors and install a water pump or a larger motor nearby and switch it on and off repeatedly. I was hoping I could replicate the issue but so far I haven't had that luck..

    Istvan
    RainMachine Support

  • Avatar
    Istvan

    Hi,

    Just following up on this matter - I tried running a longer piece of wire all around a garage door opener motor and made different setups by plugging the cable in the electrical nets in question, but couldn't trigger this behavior yet. I will be making more experiments and share my findings.

    How does the resistor setup behave on your system?

    Istvan
    RainMachine Support

Please sign in to leave a comment.