Over-designed salt water exchange system

Discussion in 'DIY' started by rygh, Mar 13, 2014.

  1. tygunn

    tygunn Webmaster

    The optical switches are pretty reliable; my Tunze Osmolator works great with it. However, the can fail; I had an instance where the osmolator would keep "seeing" low water level and trigger a short top off until it would eventually hit the backup float switch.

    Interesting idea of using the linear potentiometer and a float to get an indication of the current water level.
    Have you considered a pressure-based sensor?
    eg. http://www.newark.com/freescale-sem...ic-pressure-sensor-0-to-10kpa-1351/dp/14R8944
    Or an ultrasonic one that is waterproof?
  2. rygh

    rygh Supporting Member

    Good data about the optical switches. I am not sure what I will do there yet.

    Yes, I considered the other methods.
    Pressure based is nice, but they can drift. Leaking is not uncommon, which totally throws it off.
    But even with no leaks, air diffuses into the water.
    Ultrasonic is really inaccurate unless you get a very expensive one.
    That one on ebay is only 1 cm accuracy.
    Although if only for monitoring, might be ok now.
  3. tygunn

    tygunn Webmaster

    I didn't think about the leaking aspect and air for the pressure sensors; interesting.

    How big of a linear pot are you going to use? Presumably the length of the pot is going to be the limiting factor for how much change in water level you can measure.

    Very interested to ssee how it'll work out for you. I've always wanted to measure water levels accurately in the tank, sump, and top off basins.
  4. rygh

    rygh Supporting Member

    I measured removing 10 gallons, and it was 2 inches of drop in the sump.
    I could use a standard 65 mm (2.6") travel slide pot.
    But I might bump it up to a 100mm one.
  5. rygh

    rygh Supporting Member

    Well, with my tank crash and overhaul, one thing I want to do is finish this project.
    I rather stalled as I ran low on time and tank had other issues.
    BUT: I will be simplifying it a bit, so it gets done.
    The main thing : No fancy analog water level measurement. Just using simple float switches.

    Here is the latest schematic.
    Basically, a little I/O expander to take I2C from RASPI for programmable outputs.
    Then a darlington driver, for higher current to the relays.
    And a few random other ports, since I may use this on other projects.

    I might make several boards, and they are pretty generic, so on the off chance that
    others are interested in RASPI based fish controllers, let me know.

    Attached Files:

    neuro likes this.
  6. rygh

    rygh Supporting Member

    Looks like the board will be here after Thanksgiving.
    Ehh, I am cheap, no hurry.
    Here is the board, with top pour layer removed for better visibility.
    Funny ... free tools are worth what you pay for them. Autorouter sucked.
    neuro and HiFidelity like this.
  7. rygh

    rygh Supporting Member

    I made a new box for float switches.
    It started out really as just a way to keep the snails and crud out.

    But an inspiration hit : Have the RODI from top-off enter through this box.
    That will keep all the marine life from growing on the switches.
    Plus should eliminate calcium buildup.

    The switches are on an insert that pulls out for easy cleaning.
    Middle one is for normal level. Top is emergency overflow kill switch.
    Bottom is for water exchange.

    Coral reefer and neuro like this.
  8. tygunn

    tygunn Webmaster

    Looks awesome! What software did you use for the circuit board? I'm starting down the path of learning basic electronics and have been playing with kicad a bit. Not sure if there are.better options.

    Sent from my Nexus 5 using Tapatalk
  9. rygh

    rygh Supporting Member

    I used PCB Creator.
    The PCB Creator tool is VERY easy to use. Simplest user interface of any I have tried.
    It is provided free by Bay Area Circuits, who I used to build the boards.
    It makes purchasing boards easy.
    A nice, local, cheap but good quality, board house.
    **BUT** It is a bit weak in the component library. Particularly connectors.
    And I made the unfortunately mistake of copying someone elses. More on that later.

    I have not heard great things about KiCad, but I have never used it myself, so no
    real opinion there.

    I absolutely HATE eagle. The user interface is simply horrible.
    But, there is a ton of hobbyist content out there. Both schematics and components.
    It is the most popular non-industry one. But I think mostly because it was the
    only free one that was not total junk, for a long time.

    IMO: The best real tool is OrCad. That is what most corporations use.
    Best if you have a team of engineers doing a serious board.
    And they do have some sort of free limited version.
    But it does take some setup and learning.
  10. rygh

    rygh Supporting Member

    So the board is back. See below.
    So far, no smoke. But I did not test much last night.

    One annoying problem though: The holes on the 8 pin connectors were too small!!
    The 2,4, and 10 pins were fine. I made those cells myself. But I copied the 8 from a sample schematic,
    and the holes were tiny. Grr. I manage to find a 16 pin dip socket with small pins lying around, so used
    that, then plugged screw-header into the socket. Will work, but GRRR. :mad:

    Now for some python coding.

  11. HiFidelity

    HiFidelity Guest

    hey Mark, what can I do to get you to teach me your ways with the electronics haha :D
  12. Coral reefer

    Coral reefer President

    What can I do to get him to speak English? Wish I understood any of what he said in the last two posts....
  13. rygh

    rygh Supporting Member

    No problem:

    Turn on, not catch on fire .... Me be happy.
    Wire no go in hole ... argh. Hit with hammer to make fit ... finger hurt, but still not fit. Me be sad.
    Now need to push clicky things on keyboard lots, make magic happen inside.
  14. Coral reefer

    Coral reefer President

    Uhhhhh! Me understand now...shoulda known it had magic keyboard stuff hapenning.
  15. rygh

    rygh Supporting Member

    So as expected, this project is going very slowly. Never enough time to work on it.

    BUT: I think I may change the project a bit.
    I had an inspiration, that with this water change system, a PH probe,
    2 extra pumps, and a bit of dilute acid, I think I can auto-test Alkalinity.
    And since my Alk notoriously likes to drift, it might be worth it.
    More on that later.

    This is the current system.
    RaspberryPi on upper left.
    My board in center with too much glare.
    Small LCD below RasPi.
    Power supply on right.
    Unnecessarily large relays next to power supply.

    All nicely mounted way too close to my RODI.
    (I think I may move that)

    HiFidelity likes this.
  16. tygunn

    tygunn Webmaster

    Explain the alk auto test, this has my curiosity going. :)

    Sent from my Nexus 5 using Tapatalk
  17. rygh

    rygh Supporting Member

    I was writing down some informal ideas on that.
    Since I doubt I will build it soon, a fancier write-up might benefit me long term as well.
    So stay tuned.
  18. tygunn

    tygunn Webmaster

    Okay, I'll stay tuned. :) Anything to get me out of doing tests manually.
  19. rygh

    rygh Supporting Member

  20. rygh

    rygh Supporting Member

    Project is back on!! :cool:
    Ok, I am pretty sure nobody else knew it was dead .... but whatever.

    What happened: Stupid (add lots of bad words here) Raspberry Pi is what happened!

    That device is great for a little web or video server, but terrible for reliable real time automation.
    On boot, it tends to rather randomly enable and toggle I/O. Really!
    They expect everyone to have displays and so on.
    The fix is to edit the kernel ...... ARGH
    Then there are the fun things that happen when internet goes down. Especially during boot.
    And the not so real time issues with linux.
    And it is very susceptible to ground noise.
    And a mess with a mix of Python and C. Well, that was a bad decision on my part.

    I got too frustrated, and basically gave up.
    I used the tanks and pumps for semi-manual water change.

    I always meant to get back to it and finish, but using an Arduino again ... boring.
    Just not interesting enough to get motivated.

    Then along came the post by Erin about the new Modulo.:D
    I could use the Particle photon, plus get some handy components around it, with little effort or cost.
    Motivated again!
    Have to give her some credit. Thanks! @pixelpixi

    Of course, still slow going, since I never have enough time.
    So maybe only another year....

    At any rate, here is the NEW plan.

    Current status: About 80% done.
    Controller works. Code is mostly done. Sensors are done. Pumps and tanks were already done.
    I decided to fab a real PCB for the relay section though, so waiting on that now.
    (A PCB done on the cheap means it will be a bit)
    jonmos75 and Enderturtle like this.

Share This Page