Coffee Auto-fill

Discussion in 'The Projects Forum' started by colbyjax, Jun 21, 2013.

  1. colbyjax

    Thread Starter New Member

    Jun 21, 2013
    7
    0
    I'm new to electronics and I'm trying to design a circuit to automatically fill a coffee pot. This would operate a water solenoid valve to fill the container. Once the container is filled to a set fill amount (one switch) or if it hits the kill switch (meaning the water would spill out, kind of a backup) the water solenoid valve would cut-off.

    A switch would turn the system on. The hard part is, that once the fill valve cuts the water solenoid off, I don't want it to come back on. So if the coffee is brewing, lowering the water level, I don't want the solenoid to come back on. I designed a switch using a latching relay, but I think this design could be simplified. Looking for ideas.

    Any help appreciated in advance.
     
  2. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    But how does the process start off? What turns on the solenoid to begin with?
     
  3. colbyjax

    Thread Starter New Member

    Jun 21, 2013
    7
    0
    Hi ErnieM,

    Thanks for getting back to me.
    When the 'System on' Switch is turned on, the relay has not been fired (until the fill switch).....so the relay contact (it is double throw) passes the voltage to the water solenoid (indicated by yellow led). When the fill switch (or kill switch) is activated, it fires the relay, so the contact breaks from the solenoid and connects to the brew process (led).

    I have a very elementary understanding here, so I took a shot at it.....I was able to build this on a breadboard. It seems to me there might be an easier way using a transistor or MOSFET, but I just don't know enough about those to make that happen.
     
  4. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    Welcome to AAC.

    I'm not quite following your drawing, but I think I understand what you're trying to do.

    Take a look at the attached flowchart. Does this accurately describe what you're trying to accomplish?

    If so, you can do with a 555 monostable circuit set for x seconds (max time you expect to fill an empty pot) with an edge-trigger and some logic gates - a 3-input AND gate and a couple of NOT gates depending on the sensors used.

    Let me know if the flowchart is correct and how you plan to detect when the pot is a) filled and b) overflowing. If possible, post a link to the sensors you intend to use.
     
  5. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    elec_mech has a good point: design your requirements first, then design a circuit to perform that function.
     
  6. colbyjax

    Thread Starter New Member

    Jun 21, 2013
    7
    0
    Ahhh.....exactly. That flowchart is right on. I probably should have shown a flowchart first. I had done something similar and took a shot at the electrical diagram. I had seen some info on using the 555 as a monostable, but I wasn't sure how. Can you post a diagram of how this would work?

    I had planned to use three float switches (two at the top of the container (fill and overflow). One at the bottom (empty).

    I am very interested in seeing how to use the 555. I considered using the arduino, but wanted to enhance my knowledge of 'raw' electronics without the programming for now.

    Float switches similar to the following:
    http://www.amazon.com/Liquid-Water-...8&qid=1371905297&sr=8-2&keywords=float+switch
     
  7. wayneh

    Expert

    Sep 9, 2010
    12,103
    3,038
    You would enjoy disassembling a broken Keurig coffee machine if you can find one, to see how they solved some of these issues. Of course it uses a computer but the physical devices - pumps, sensors, etc. - are fascinating. These things are known for failing due to the PCB which cannot be replaced. All the mechanicals keep working fine but the machine heads to the dump because the board goes out. :mad:
     
  8. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    I had two broken ones I attempted to fix, but it was more trouble disassembling them than it was worth. I had to look online until I found someone who gave a small written dissertation on how to do so and gave up. I've seen numerous problems with these ranging from leaking pumps to failed transformers. I agree though, these are neat devices, just not made for easy fixing.

    Okay, I've put together a timing diagram, logic diagram, and schematic. Can't promise it's error-free, but should be close. I did not include an empty sensor, but you can add one if you really want. Currently though, there are two redundant checks being made - one is performed by the timing of the 555, the other is with the overflow sensor.

    This project does lend itself well to a microcontroller solution, but like yourself, I like exploring the hardwiring option when it is available.
     
  9. BillO

    Well-Known Member

    Nov 24, 2008
    985
    136
    I would not rely on a timer for filling as water pressure can fluctuate such that the fill time may vary quite a lot, so you would need to depend on the level sensors anyway. Not to detract from elec mech's fine effort.

    Almost any relay can be used for this and it can be kept very, very simple. Have a look at the schematic.

    If the tank is below the fill level, power will be available to the relay's contacts and SW1. Pushing SW1 will cause the relay to latch until either SW2 or SW3 open. When either SW2 or SW3 open, power is removed from the relay which then tuns off and unlatches. The solenoid valve is across the relay, so it follows the relay's operation
     
    elec_mech likes this.
  10. wayneh

    Expert

    Sep 9, 2010
    12,103
    3,038
    I could be wrong but I think timed fill is exactly how the Keurig works, although probably with backup sensors for overflow or such. Depending on the pump, timed delivery can be very accurate. It's how a lot of laboratory dispensing pumps work, for instance.
     
  11. BillO

    Well-Known Member

    Nov 24, 2008
    985
    136
    Can't argue with that. Fixed displacement pumps can and do deliver accurate volumes over time. I was referring to depending on household/ambient water pressure, which can vary substantially. Ever been in the shower when someone flushes the toilet?

    The Keurig systems I know certainly use a pump to pass the hot water through the cup containing the coffee grounds, but I'm not sure they fill their reservoirs using a pump. At least not the little one I have at home, or the big fancy one we had at work. However, there are so many different models, and I have no idea how they all work.
     
  12. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    Thank you. :D I only included the timer as a redundant safety mechanism however, not as the primary means to determine how much to fill the container. As I inferred from the OP, the water needed to fill the container may vary from time to time. I was more thinking that if all else failed, namely the water sensors, the water would shut off after a preset amount of time. This wouldn't necessarily prevent spillover, but it would at least cut the water off after a set time.

    BillO, is the relay you're referring to a latching relay? I didn't see an SCR or similar circuitry, but just wanted to verify. It is a simple and elegant solution, kudos. :)
     
    ErnieM likes this.
  13. BillO

    Well-Known Member

    Nov 24, 2008
    985
    136
    It's just a regular relay, but wired to latch. SW1 supplies it with +V when pushed, then once it pulls in, it supplies +V itself so latches until either SW2 or SW3 open.

    Gotcha.
     
  14. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    Ahh, I see it now. My hat's off to you.
     
  15. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    +1 That's what I thought your intent was.

    The last thing you want is an ever-running spigot if (when) the FULL sensor bricks.
     
  16. #12

    Expert

    Nov 30, 2010
    16,278
    6,790
    Every time I see a question like this, it makes me think I want a cup of coffee:mad:

    I'm going to do it the hard way. I'll walk to the kitchen and put water in a container, then heat it.
     
  17. colbyjax

    Thread Starter New Member

    Jun 21, 2013
    7
    0
    Wow thanks for the replies.
    I am going over elec_mechs design in detail trying to understand it. I will have some follow-up questions on that.

    So with BillO's design, once the water is full, it shuts off the water solenoid. But once the water starts to go through the brew process, and the water level drops, wouldn't it just trigger the solenoid to start again (creating an infinite brew)?

    Or am I missing something there?

    Big thanks for the replies btw.....learning a lot here.
     
  18. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    No, BillO's design works very similar to mine - the whole process only works once with each press of the button. Once the solenoid shuts off, the power to the circuit is cut off completely and will only operate if the button is pressed again.
     
  19. colbyjax

    Thread Starter New Member

    Jun 21, 2013
    7
    0
    Gotcha, I see that now. Pretty simple, elegant design.

    I am looking over elec_mech diagram now. I am trying to figure it out to increase my understanding of circuitry.

    I think I understand the 555, it sends a timed pulse (out of the #3 pin) to the CD4082.

    This is where I'm a little confused.

    Does the fill sensor send voltage to #2 pin and overflow sensor send voltage to #6 pin? If so, is this a NOR to block #4 from sending voltage?

    I'm just thinking that the logic is..... 2 OR 6 = 1 then 4 is 0?

    These questions are probably elementary, but I really appreciate the help. Also, what is the MOSFET for?

    Thanks again guys.
     
  20. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    Oops, U2 should be a CD4081, not 4082. It is a quad 2-input AND gate. Take a look at the 4081 functional diagram here.

    I'm assuming your water level sensors are essentially normally open switches. Thus, one side is pulled high and the other is connected directly to ground. The high side is connected to AND gate inputs. If both sensors are not tripped, then the input to the AND gate is high from each sensor.

    The output from one AND gate is fed to the input of another making a 3-input AND gate. If all three inputs are high, then the output is high which turns on the MOSFET and turns on the water solenoid.

    If the 555 is off, the output is low. If both sensors are not tripped, their outputs to the AND gates are high. Because this is a 3-input AND gate and the 555 output is low, the AND gate output is low and the water is off.

    Once the 555 is turned on, it outputs a high signal for up to 100 seconds - you can change this by varying the potentiometer. Once the 555 output goes high, the 3 inputs to the AND are now all high and output of the AND goes high turning on the water.

    The water will stay on until one of the sensors is tripped (causing one of the AND gate inputs to go low) or the 555 timing cycle completes and the 555 output goes low.

    The AND gate cannot handle much more than a few mA and a solenoid typically requires hundreds of mA or more, hence the need for a MOSFET. The MOSFET is effectively acting as a relay while drawing next to no current from the logic circuit.

    Let me know if you have any other questions or anything I've explained thus far is unclear.
     
Loading...