Newbie needing direction on low water level alarm circuit

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

  1. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    Munch away.

    At one point, I mentioned a header. This http://www.digikey.com/product-detail/en/aries-electronics/16-600-10/A103-ND/4260 is what I was talking about.

    The idea would be:

    1. to use these for the input/output of the comparitor. For example you could try out different filter combinations and different valued resistors without wasting your main board. They are re-useable to a point, but the hand made breadboard even less so. You can do things to help out in that area, by using 0.042 hole push in pins by Vector.

    With your comparitor on a DIP socket, you could swap out an inverting comparitor if you liked.
    These headers would be socketed.

    Doing stuff on the output, maybe you could use an LED for trial #1 and something else for trial #2. ANother part, that could be useful is this one: http://www.clare.com/home/pdfs.nsf/0/B850B566672E5B828825806F004CB226/$file/CPC2907B.pdf or similar.

    You could have one section operate the LED with the cr2032 battery and another use an entirely different power supply.

    But, i still like playing with the input/filter circuit with a LED and then graduate.

    Search out buzzers and see if they are loud enough at 3V, 9V, 12V whatever.

    Here: http://www.mallory-sonalert.com/AVBySeries.aspx there are buzzers that may not even be suitable, but you can listen to their sounds etc, but probably won't be able to tell how loud they are. I put a sonalert on my "alarm clock" which is actually a timer. In order to guarantee that I wake-up, I have to turn the clock around so the sounder is facing me. I did use the variable volume disc.

    For like fun and giggles, you could use a wall wart for the alarm buzzer, but make make the probes use a battery for safety reasons. These are just possibilities. Thrown out as an idea, not advocated.
     
  2. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    Hi KeepIt! I'm embarrassed that I have been away for so long. One of the problems is I'm doing this on the side and I don't have a lot of time, but in reality, I think the bigger issue is that I'm pretty lost. You are very helpful to suggest options, but when those options are full of acronyms and terminology I don't understand, I go down a google rabbit hole and still end up lost! lol. Maybe this stuff is harder to figure out than I realized. On the other hand, I really am less clueless than when we started this.

    I have an idea, though, if you are still reading this thread. I'd like to move ahead with buying the parts that will allow me to start learning how to make this thing work. I thought that once we knew the voltage output from the broth cell, that the rest would be more or less boilerplate. I've come to learn complications like we don't know how many mA are generated by the cell, and therefore might damage the comparator (hence the need for 200 ohm resistor in series, if I read you right).

    So would it be possible for me to provide you with some requirements, and then you recommend a parts list for me to get started with? All of these components seem to be only a matter of pennies and dollars so I don't mind buying things I don't really need. With that in mind, let me recap where I think we are at, and make a wish list:

    1. I think it is important that the electronics/battery be table mounted, with wires going to the probe mounted at the pot. It is just too hot and humid to try and mount electronics right above a hot pot over an open flame. Ultimately I think the probe wires would have to be around 3' long.
    2. I don't want the final product to go through batteries too quickly. I'm thinking the 9v battery is best if it will maximize battery life for powering an alarm. I assume we will still use a lower power battery for use with the comparator, and that the 9v is not suitable for that. (Or can it be made suitable to work so that we only have 1 battery to worry about)? It is important that the user know that the battery is good so they don't ruin their stock because the battery died in the middle of the boil down. I hope that simply holding the probe in the air, and checking for an alarm, would suffice.
    3. Regarding how to deal with the signal as the liquid level gets to the bottom of the probe, whatever is simpler is fine. If it is easier to use a filter in front of the comparator to provide a clean on or off signal, or to use some kind of timer to indicate when the signal has been 0 mV for 30 seconds, either is fine. The liquid level does not have to be exact. I think if both options are of equal ease, then I'd choose the one that puts the bottom of the probe closest to the actual liquid level. I imagine that would be the filter.

    So given that, do we have enough information to put together some components? I think what we are looking at is: filter, 200 ohm resistor (the filter doesn't act as a resistor?), 10 M resistor, comparator, battery for comparator, LED for testing purposes, 9V battery for buzzer, header and any other parts needed to test the circuit. For now I could just use a couple of wires and twist them to the header as an input. I'm pretty sure I've got something wrong or am missing something, but I really want to stop losing time on this project. If you can confirm an order list for me, and maybe some help as to sourcing, I will get everything on order.

    Thanks for reading, and hopefully I can start putting a circuit together in the next week or two.
     
  3. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    I'm still here. Barely. Designing stuff is a real pain and I too am knee deep in alligators, so to speak.

    I'd actually like to start smaller. No buzzer - just a simple LED.

    So, we need to find a p/n that www.proto-advantage.com and digikey can work with the different senses (inverted/non-inverted in multiple pair combinations and a single output. The comparator came in single and duals, but digikey didn't stock one of them. You don't at this point want to solder them. Let them do it to an adapter. I took a peak today and got frustrated. Might as well get both versions made into adapters.

    The CR2032 battery is essentially a regulated power supply. Good for now.

    Short leads are good. The long ones may have issues like you touch the wire insulation, the alarm goes off.

    This test case, might actually even have a box (possibly sealed), but not 100%); have a way to reverse the polarity of the probes with a switch and have a Buffer for you to monitor externally what the voltage is, without really disturbing it.

    Not sure if you actually want to build a simple current meter?

    There is shielding and guarding for long leads and twisted pair shielded.

    Wiring for "foods" and high temperature is another area. PUR insulation http://sickusablog.com/pur-cables-pvc-cables/ may be required. Maybe even PTFE. FNRC is probably another. I did do one stint with FAP (Fire Alarm Panel) cable. That stuff was really stiff and expensive and difficult to get in a small quantity and EXPENSIVE.

    Small waterproof cases https://www.polycase.com/waterproof-enclosures , a possibility.
     
  4. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    I don't think I need any special cables for the sake of food contact. With the clip I've got for hanging on the pot, the wires won't be near or even over the food. There could be an issue of heat with the flame exhaust, but that's something I can test easily down the road.

    Is an enclosure needed at this point? I'm happy to buy one but I was thinking something more along the line of plastics rather than aluminum, that would still be water tight. But, at this point, it isn't critical. I don't think this needs to be water submersible just because it is in a kitchen, although if that were a cheap option then why not. Plenty of other tools in the kitchen are not supposed to get wet.

    Let me know what you think whenever you get some free time. God knows I've taken forever to get to this point.

    Dan
     
  5. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    For the most part, I agree with you.

    It's just that long cables of high impeadance do weird things. If I attach a wire to an oscilloscope with a 1 Meg Input impeadance, it picks up a voltage. I just took a piece of wire to one of my scope's inputs and without even touching it, the peak to peak voltage measured was 20 mV. Touching the insulation, I got around 120 mV.

    The line frequency is all around us and the body picks that up and can capacitively couple that to the input of the scope.

    This would be more like EMI (Electromagnetic Interference) rather than RFI (Radio Frequency Interference), so that's the real problem with long "leads". In general, "shielding" (enclosing the conductor with a ground) reduces RFI and twisting reduces EMI. The shield works because the RF can't penetrate the grounded shield. Sources of RFI would be a close radio station. The twisting works on a "differential" input because what's conducted in one wire is conducted in the opposite direction in the other wire. Thus, long wires to a "box" may cause issues.

    The comparator by itself is "single-ended" and won't necessarily reject signals common to both inputs, but twisting would help.

    A lot of this http://www.atom.fysik.lth.se/QI/las...ompeter Electronic systems wiring & cable.pdf stuff isn't relevant, but it covers the basics in great detail.

    If you were trying to measure currents on the order of nano-amps (nA) an entirely different set of circumstances arise.

    So, long leads can be an issue and that's why most of us in this thread were thinking something at the pot. Although, I don't like it, I'm thinking of a hybrid, so the wires can stay short and the buzzer/(larger battery) can be further away. The wires becomes more of a "switch".

    Sensor "wires" in mid-air MAY create an ON all the time condition without ways to tame the interference. Short leads is the first step.
     
  6. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    Thanks for the link. I'll read up on that. In my initial tests, I used the test leads from the voltmeter as the wires or leads we are talking about. I connected one lead to the pot and the other to the stainless rod, initially holding them by hand with electrical tape. More recently, I purchased several different types and gauges of wire from Home Depot and also got a set of alligator clip leads. I had all sorts of combinations of wires between the pot and the voltmeter. I had the voltmeter test lead attached to an alligator clip wire, which was then clipped to wire from HD, which was finally clipped to another wire that was clipped to the pot. So I had four kinds of wires in series. I only did this so that I could use alligator clips and switch between wires easily. Maybe this wasn't the best way to do it, but I did get pretty much the same voltages no matter which wire combinations I used. I understand this might not address the interference issues with an actual circuit, but I wanted to remind you of the above in case it sheds any light on the likelihood of there being a problem.

    When you say "short leads" are we talking less than 12", or something else?

    I originally wanted a self-contained device above the pot that was sealed, but between the steam coming off the pot and the hot exhaust from the open flame, it seemed difficult to make a battery work in such conditions (crazy thought, but maybe even mount everything above the pot including the buzzer and have just the battery on the counter top away from the heat? something to think about). I thought having a remote circuit just the way they do it with oven temperature probes would be easier. By the way, the type probe I am talking about uses a wire wrapped in something like a metal fabric. Something like this:

    http://images.cookingforengineers.com/pics6/thermapen/oventhermometer_640.jpg
     
  7. paulyvee

    New Member

    Jan 18, 2017
    2
    0
    As far as first thoughts on this I suggest fluid depth measurement would always be a bit problematical due to the ever present steam condensing around the end of the probe. I suggest a temperature probe set at the requisite depth in the fluid would see either a sharp rise when changing from water to steam as the fluid drops below the probe or a sharp temperature fall if the fluid was very dense and required higher than 100 deg centigrade to boil. This could be amplified and used as an alarm trigger. Foodgrade thermometers are available for measuring meat and suchlike and also for measuring the temperature involved in making jams - probably a much wider temperature range.
     
  8. paulyvee

    New Member

    Jan 18, 2017
    2
    0
    By the way, steam temperature MUST be= higher than boiling water temperature. At normal atmospheric pressure water temperature is limited to 100 deg C. The energy fed into the water by the heat source must go somewhere so goes to energise the water atoms sufficientyl that thyey turn into an invisible gas - steam. This steam layer them leaves the rises above the surface of the water, starts to condense and forms water vapour. This is visible but the layer of steam is not. This can be seen if you look at the spout of a kettle. At the spout itself there is no visible cloud. This is the steam area. Further away from the spout comes the boiling water vapour area. AFTER the small steam gap.
     
  9. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    That's not quite correct. You are confusing sensible heat and latent heat -- just google them. The temp of the steam just above the water level is the same as the water.

    We've considered using a temperature probe but there are issues with that as well. I still think we have the easiest solution, which is a plain stainless rod - cheap and hard to break. If for some reason the circuit is difficult doing it this way then I'll take another look at temp probes.

    Thanks for your comments!
     
  10. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    Just like you can have water a 32F, you can have ice at 32F. The same goes for steam. The impurities may change the temperature of boiling because it's no longer water, so temperature isn't viable unless it's self-cooling of thermister sensor.

    The probe length counts as leads.

    Most voltmeters use the dual-slope integration technique. See: https://www.maximintegrated.com/en/app-notes/index.mvp/id/1041 and actually do the integration at the power line frequency, thus the effects I'm mentioning inherently go away. I'm remembering the "cheap" aspect.

    So, yea, a complex design using a processor and integrating A/D converter is a design that has a better chance of working.

    That said...

    What I'd actually like you to do is to pick some resistors in the range of 1K, 10K, 100K, 1M and 10 M, infinity and put them in parallel with your meter while in the soup mix and look at the changes in voltage and come up with a table.

    ------------------Req____V...........Soup (pick one or two- high and low voltage)
    ....inf||10M.....10M
    .10M||10M ..... 5M
    100K||10M......99.09K
    etc.

    The second 10 M is the input Z of your meter. The Req is the equivalent resistance 1/Req=1/R1+1/R2 or the parallel combination.

    While your measuring a voltage, your also measuring a current. A bit small, but still a current.
    So if your measuring 100 mV, the current is I=100e-3/10e+6 or about a uA.

    With at least 2 equations and two unknowns, I should be able to calculate the resistance of the "soup" which, I think, would help.

    I just picked an IA Instrumentation AMP) http://www.analog.com/media/en/technical-documentation/data-sheets/AD8237.pdf out of a hat. Like I said earlier, they are limited in supply range 1.8 to 5.5V in this case.

    This configuration "opens up" the ability to "add filters", rejects common-mode signals, allows +- potential , probably makes it easier to have it further away (long wires).

    So, call this the "complex case", but it has it's own issues.

    And basically "loud buzzer" and battery powered with long life don't go together. I have an "extra loud" timer from Thermoworks (the wire example you showed) and it is loud. It uses a 9V battery. I also have another timer from them as well that uses a CR2032, I think, that hangs around your neck and I can barely hear it. A buzzer can't really be located in the presence of steam. Low voltage typically means low power.
     
  11. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    As far as the buzzer, I really don' think it has to be excessively loud. I have other digital devices like timers and a microwave alarm in the kitchen, and even with the convection oven and exhaust hood running, you can hear them. I think it is the high pitch of the digital type alarms that cuts through the other noise.

    Concerning the resistors, can I use my leads with alligator clips on either end? I would clip to the voltmeter lead and then to the resistor, and then clip another wire from the resistor to my device. Same on the other terminal. Too many wires, or would this serve your purpose? Also, I assume the "infinity" resistor is no resistor at all?

    I can also envision a configuration where the CR battery is above the pot with the electronics housed in a sealed plastic device, while the buzzer and the larger battery are on the counter. I still prefer having everything on the counter top in a box, but it is something to consider.
     
  12. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
  13. wayneh

    Expert

    Sep 9, 2010
    13,101
    3,899
    1/4W resistors are pretty much standard for low power signals. They dominate the typical workbench stuff. The only reason I'd go larger is when you're actually handling power instead of information. Then you make the calculation of the expected current and power dissipation, and get a resistor rated to twice that.
     
  14. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    Thanks, wayneh. I'll get the 1/4W.
     
  15. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    We're on the same page. For breadboarding, it sort of becomes "what to stock". The larger resistors (1/4 and 1/2 W) are easy to handle.

    Seat of the pants 10V and 20 mA is 0.2 W (P=VI) which are much bigger than expected numbers.

    For my purpose above, you can stuff a 1/8 W (leads are smaller diameter) resistor in the leads at the meter.

    At Radio Shack, you won't have all of the choices. Right now, I don't even know where one is.

    This configuration is one way of dealing with the lead length, low sensor power and buzzer beep.
    The lead pick-up is an anticipated issue.

    I can't hear at 2600 Hz.
     
  16. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    There's a Radio Shack nearby and they have a variety pack that has all the resistors on your list. They have 1/4 and 1/2 watt, so I'll pick up the 1/4 watt and see what happens.
     
  17. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    The Radio Shack nearby has the 1/8 watt resistors so I got an assorted pack of those. The wire is very thin. Do these get inserted directly into where the red and black leads would normally go, and then I use alligator clips from there? I'm not sure I understand as the wire is thin and the female port on the voltmeter is large.

    I can also pick up 1/4 watt later today if I'll be needing those later. The also have led's and breadboards. I took some pics to upload tonight. If there's something I can use that would be good as everything is on sale (they are going out of biz).
     
  18. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    I just tried a 1/2 watt resistor lead and a test lead together and it works. The 1/8 W should be easier. So, just insert the resistor in the DVM hole and put the banana plug in too which will hold the resistor in place. So, unless you have pin plugs, it should work.

    You can also put it at your alligator clips. It doesn't matter for these tests.
     
  19. danielj_2016

    Thread Starter Member

    Jun 24, 2016
    85
    3
    So I've got 5 9V batteries and the stupid voltmeter won't work with any of them. Cheap crap I guess. I'll report back when I have some data. I'm going to need an eyeglass prescription when I'm done reading the colored stripes to figure out what these resistors are. I'm trying to just use the voltmeter to confirm my color matching. Is that a green or blue stripe?
     
  20. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,755
    294
    Well, make sure the DVM is set to ohms.

    Warning... There are 3 and 4 band resistors excluding the tolerance band. 5% gold, 10% Silver and no band 20%

    So, some resistors use 2 digits and a multiplier. Others use 3 digits and a multiplier.

    Eventually, you get to the point that you can pick up the resistor and instantly know it's value.
     
Loading...