Newbie needing direction on low water level alarm circuit

Discussion in 'The Projects Forum' started by danielj_2016, Jun 24, 2016.

  1. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    Hi folks This looks like a great forum! I've been reading some background material and trying to see if the answer to my question is in the forum, but I haven't found anything. Let me say briefly that I "used to be" a chemical engineer, so I have some technical background, but it was CHEMICAL. I'm pretty lost when it comes to electronics. I understand basic generation a/c and d/c, circuits in school 30 years ago, etc. What I know about circuit boards would fill 1/2 a paragraph, lol.

    I am trying to design what I think is a simple device to sound an audible alarm when boiling water in a pot reaches a low level. I wanted to achieve this without a mechanical float, and I found a circuit that seems to fit the bill, sort of. Below is a copy of the circuit, but the designer at circuitsgallery.com doesn't seem interested in helping people figure out why it doesn't often work. The principle, as I understand it, is that a probe is inserted into the water do the desired depth. This provides a ground and keeps a transistor closed (learned about how transistors work last night, lol). When the water drops below that level, the voltage in the line opens the transistor and causes the alarm.

    I have a number questions, but I'd like to ask just a couple to get started (don't want to scare anybody off with 1000 questions about something so simple).

    1. Would you consider this generally a good design, and one that could be modified to work for my specific situation. For instance, if I want to use an 18" long stainless rod as the probe, I imagine I might have to change some of the components in the circuit in order to make it work.

    2. Can this be made to work using a battery rather than a/c power supply? I would like to make it portable and to use as little battery power as possible.

    Hopefully I've described things adequately. Any help, or direction if this is the wrong forum would be greatly appreciated!

    Dan
     
  2. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,977
    744
    Your circuit will work on a 6v battery ok, i would use a Darlington pair transistors or mosfet, the only problem would be erosion of the electrodes.
     
  3. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    I was hoping the current to the water would be so small as to be unmeasureable, but if there is any hydrolysis going on, I can't use this design.
     
  4. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,977
    744
    You're better using an ac signal or better still a float sensor.
     
  5. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    Actually a small amount of electrolytic reaction is not a problem. The problem is degradation of the sensor. I'm thinking if the probe isn't too long I could conceivably use a platinum probe. That would be inert as far as taking part in an electrolytic reaction.

    Could this be done with battery power? I don't think the project is really practical if I need AC. Also, I don't want to give up on the idea of a sensor with no moving parts like this. I could use a float, but I really want to exhaust any options with this simple design, if I can make it work.
     
  6. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    Project : Pet Water Dish Alarm Sensor

    Keep the current and power down in the micro region and probe decay is not a problem.

    The project above runs for many months continuously, I have had to change the probes because they got rusty and months of continuous use, but they have not failed. Simple iron or zinc coated wire works fine, and I do not believe they would be toxic.
     
  7. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,977
    744
    yes you can do it with a battery, but you need to drop the current down to microamps in the probes, so a mosfet transistor sensor would be best
     
  8. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,779
    1,103
    How pure is the water? Does it have a high mineral content?
    Pure water is not very conductive, so for this the probe circuit posted would be unlikely to work reliably.
     
    PAFMC01 likes this.
  9. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    Thanks for the replies. Wendy, thanks for posting that. I'll need some direction if I try to reproduce it. :)
    I don't think I would need that 7555 chip, if that's what it is, because I only need a buzzer to sound.

    I'm using this in a kitchen for reducing liquids like chicken broth. I might start boiling 12" worth of dilute broth down to say 3" of concentrate. It is easy to overshoot because it takes time to boil all the water out, and having an alarm would help a lot. So conductivity isn't a problem. However, I can't have any metals dissolving into the liquid. It typically might take 2 to 4 hours to boil down, so the probe will be in boiling water for hours at a time, but not 24/7. Basically, it has to be food safe. I might be able to use a plastic coated wire with a platinum tip, if I need to go that far. Also, foaming might be a problem but I think that might be solvable if I can get this first part to work right.
     
  10. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,779
    1,103
    Hmm, using the probe in boiling liquid complicates things because even when the liquid drops below the probe tip level the steam may well be conductive enough to fool the circuit. I suspect it's going to need some experimentation to find optimum resistor values.
     
  11. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    Yes, that's the kind of thing I was getting to, as well. For instance, does the size/length of the probe dictate what the other components need to be (larger/smaller resistors, etc.). In other words, it seems to me the whole circuit needs to be in some form of balance in order for the transistor to open and do its job. Of course experimentation is often the best route, so I'm up for trying things to see what happens.
     
  12. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,523
    1,247
    The circuit in post #1 might have a problem, depending on the liquid. As shown, there must be about 4.5 mA current through the liquid to hold off the transistor. That's a lot. You can help us here by measuring the DC resistance of some of the liquids you want to sense, at various concentrations. Just stick two ohmmeter probes into the soup about 1 inch apart. Repeat at 6" apart. This will give us the "sensor" resistance range for the high state.

    Next comes setting a stable decision point. This has two components, the circuit and the mechanical interactions between the liquid and the probes.

    For the circuit, a simple 1 (post #1) or 2 (Wendy's link) transistor voltage sensor is maybe too simple. Better is a slow/cheap/stupid single-supply opamp acting as a comparator. Plenty to choose from, and CMOS flavors draw next to nothing. This will give you a nice clean snappy signal to do something with. Also, the sense current can be so low that corrosion is not a problem.

    For the soup, this is like the dryness sensor in my dryer. It is two contact discs on the wall that the clothes rub against. They go to a very high impedance comparator circuit with about a 1 to 5 minute time constant. When the clothes are dry enough that no damp spots have completed the circuit between the contacts for 5 minutes, the dryer pronounces the cycle complete. Bubbling soup seems like a similar situation - constant contact (immersed), then intermittent, then less intermittent, then a decision is made. Of course a 5 minute time constant is too long for what you are trying to control, but I think this covers the idea.

    ak
     
  13. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    Thanks, AK. I can take the resistance measurements you suggested and post them here.

    Let me try to interpret the rest and see if it got it right: Although the Darlington or mosfet transistor would work to boost the weakest possible signal we can create, it will not address the problem of an intermittent result based on foaming or bubbling liquid. To address this, you suggest a CMOS chip which is like a smarter transistor because we can program it to provide a steady signal to the alarm right when it starts to sense foaming, or even when it no longer senses foaming (like your dryer example) depending on what we want. Do I have that right?

    Also, will I still need one battery for the CMOS and another to power the alarm?
     
  14. wayneh

    Expert

    Sep 9, 2010
    12,100
    3,034
    I'm thinking there has to be a better way. I don't know why you would resist using a float. I think it's probably the most reliable and cleanest approach. Maybe a pin-pong ball sphere of stainless steel that pulls on a chain outside the pot when it has dropped far enough.

    Or, consider adding an agitator (which might help the process anyway?). When the blades are clear of the liquid, the motor load will drop and the rpm may rise. This could be detected fairly easily.

    A thermal patch on the outside of the pot might be able to sense the level. And then there are capacitive sensors that work from the outside (on metal pots?).

    Conductivity would probably be my last choice.
     
  15. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,523
    1,247
    I agree that there might be better ways to do this, but for now I'm just following along.

    Here is how a comparator improves the circuit in post #1. In round numbers, that is an on/off circuit. Liquid contact = transistor off. No contact = transistor on, sinking current to light a light or ring a bell or whatever. In that situation, surface churning means intermittent contact with the probe, so the transistor is rapidly turning on and off like yanking up and down on a light switch very fast. With practice, you can position the probe such that when the liquid stops all contact it is at the level you want, but electronic circuits don't think that way. Also, the buzzer has been going on and off rapidly for 15 minutes and your staff has quit. Is a signal that was bouncing and now is stable, stable because it really is stable, or is this just an extra long period between bounces? This is the central question to all kinds of sensor circuits. My dryer has a 5 minute decision window. If there is a valid damp contact at 4:59, the decision timer resets to zero and starts counting again. There are analog circuits for doing this called a peak detector or lowpass filter or missing pulse detector. All are based on a decision level time period.

    Separate from that is the amount of contact between the probe and the liquid. As the liquid boils off and the level decreases, the total contact area between the probe and the liquid decreases. So while there still is constant contact, the signal from the probe is changing. How a circuit deals with this depends on its design. A single mosfet has a very "soft" or rounded knee in its response curve, and the knee value (ok, pros, bear with me here) can vary by over a volt among identical devices; a bipolar transistor has a much better defined knee both in terms of its absolute value and the sharpness of the transition from off to on. But the best way is with a voltage comparator. This circuit senses the difference between two voltages down to millivolts or even microvolts (not needed here), and is setable with two resistors.

    Its output still has the bouncing problem for liquid that is churning, but the bouncing is much more repeatable. To the output of the comparator you add a resistor-capacitor filter that converts the bouncing to a nominal DC voltage. As the liquid gets lower there is less bouncing, and the DC signal decreases. A second comparator (there are two in a package) tests this slowly varying DC level against another programmed voltage, and decides when the bouncing is slow enough to proclaim that the reduction is done.

    Lotta words, actually a pretty simple circuit, identical to what determines if the fan in your PC is turning fast enough or if its bearings have started to fail.

    ak
     
  16. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    @ak: OK I think I understand the gist of what you are saying. If I get some resistance measurements from the stock (soup if you will) and it seems like this probe design is appropriate, I'm hoping you guys/gals can help me with the circuit so I can look at each component and get an understanding of what it is doing and why. I understand what you are saying about it sounding more complicated than it is. However, it does sound like a good trick to be able to deal with the intermittent signal successfully.

    @wayneh: Thanks for the input. I am open to any design, really. My second choice is probably a float. However, there are some factors to be aware of, which led me to this particular design. First, this is a commercial kitchen. Things get used... a lot. In my case only 1 or 2 careful people would be using this device, but not everybody I know who might use this instrument is so careful. So, it is attractive to have something with no delicate parts, ie moving parts (other than the circuit of course). This design is basically clamping a rod against the stock pot and setting the depth, turn on the electronics and you're done. I like the simplicity if it can be made to work. Actually locating the electronic board is another issue.

    Other issues:
    - my burners are 32,000 BTU each. That's a lot of flame. Anything on or near the outside of the pot risks getting overheated by the hot combustion coming off the open flame.

    - Anything that moves inside the liquid like a float or a spinning paddle also can get gummed up. These liquids mostly contain organic compounds (chicken broth, for example) and stuff like that gets into every nook and cranny. The easier it is to clean, the longer it will last before getting broken.

    - Also, I need this to be battery powered with as little juice as possible, so powering something to rotate is probably out.

    So those are some of the other considerations. I'm sure you've seen oven thermometers where you stick a probe into a roast and the cord comes out of the oven to an electronic thermometer. Easy and foolproof. I'm looking for something like that and this probe idea seems to fit the bill. Of course I'm always open to other perspectives.
     
    wayneh likes this.
  17. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    OK, I tried to get some resistance measurements. I don't know if my battery is OK (don't see any low battery indication) but the ohm readings bounce around a lot. Maybe this is just because I am reading very low levels of resistance?

    When I touch the leads together I get about 40 ohm (I am using the lowest range of 0 - 200 ohm). However, if I shut the multi meter off and do it again I might get 15 ohm. In any case, I took the two extremes. In one case I measured a final, thick gravy at 1/2 and 6 inches and it came in at 88 ohm both times. Pure water from the tap, which is what we start with in a chicken broth, was 125 ohm at 1/2" and 140 ohm at 6". However, when I used cold water I got 175 and 200 respectively. All these numbers bounce around quite a bit and I'm wondering if I am basically so low on resistance that I am pretty much measuring noise, of if I need a new battery.
     
  18. wayneh

    Expert

    Sep 9, 2010
    12,100
    3,034
    That's really poor. Get yourself a free meter from Harbor Freight if/when you can. Any decent meter should read very close to zero with the leads touching, and should be repeatable. You might also try your meter on a few resistors of known resistance in the range you're using, eg. 47-180Ω or whatever you have.
    I'm surprised there isn't a bigger difference between tap water and concentrated "gravy". It seems you may be able to see the difference between softened (hot) water and unsoftened (cold) water. In my home, the cold water at the kitchen tap is not softened. Maybe that's not the case in your kitchen.

    I'd be reluctant to read much into your results until you resolve the issues with your meter.
     
  19. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    41
    2
    None of the water is softened here. I thought there might be a resistance difference between boiling and cold water, but it seems I have multimeter issues. It seems to read voltage just fine, but maybe it is a cheapo unit. It is a MAS830B from Commercial Electric.
     
  20. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,523
    1,247
    A low battery shows up first as errors in the ohms ranges, because besides powering the meter, the battery has to pump current through the external resistance being measured.

    ak
     
Loading...