A vibrating object is another possibility, like an ultrasonic membrane or a piezoelectric disk

I think that would be practical. The presence/absence of liquid would produce an amplitude and/or frequency shift of the vibration. The vibrating probe could be made liquid-tight and easily cleanable.

 

Some other ideas: http://www.site.uottawa.ca/~rhabash/ELG4135L8.pdf

But, a negative peak detector. These operate off a dual supply.

 

Somehow I missed these replies. If it is becoming cumbersome to design something that will work with a single probe, maybe some of the alternatives would make life easier. I liked the probe because it was simple/cheap/hard to break. The drawback to the probe was the potential electrolytic reaction and problems that might cause, plus it sounds tricky to get a good signal, if I read you guys right.

So alternatives listed look like IR or vibration (membrane or piezeoelectric). Assuming either could be made easy to clean and NOT FRAGILE, which consumes the least power? I'd like to run this off a 9v battery. Would the electrical requirements be much greater for one of these "detectors" compared to the single probe idea? (Single probe is still my preference, but if it isn't realistic then I could go with the above alternatives).

 

Yes, powering something to vibrate or running an LED requires more power than a conductivity meter.

Have you experimented with one of those soil moisture detectors? There are lots of DIY projects to build them but you can also just buy one for not a lot of money. It might prove out the concept.

 

Somehow I missed these replies. If it is becoming cumbersome to design something that will work with a single probe, maybe some of the alternatives would make life easier. I liked the probe because it was simple/cheap/hard to break. The drawback to the probe was the potential electrolytic reaction and problems that might cause, plus it sounds tricky to get a good signal, if I read you guys right.

You still have two probes. The pot and the probe. And you have to make a connection to the pot.

Two probes encircled with a perforated tube does make a self-contained probe, but it might be harder to clean. But you may need to keep the "big parts" like bones away from the probes and that would be what the perforated tube would do.

Without spending time designing on your own from total scratch, this http://www.tauntek.com/ company has some unique products. TinyIR and the mimic series. So, take any remote, learn the code and be able to receive it.

It MAY be possible to run from a CR2032 battery for at least the pot side. The alarm side might be another story. Two part has some advantages and so does one. Using an I-V converter will very likely be lower power.

So alternatives listed look like IR or vibration (membrane or piezeoelectric). Assuming either could be made easy to clean and NOT FRAGILE, which consumes the least power? I'd like to run this off a 9v battery. Would the electrical requirements be much greater for one of these "detectors" compared to the single probe idea? (Single probe is still my preference, but if it isn't realistic then I could go with the above alternatives).

IR would be power hungry.

A "stupid" photo transistor https://www.elprocus.com/phototransistor-basics-and-advantages/ ; http://optekinc.com/pdf/App Bulletin 213-Opto Components.pdf MIGHT even work IF the sensor can see high temperature.

This idea being, assume there is a light (ambient) that the pot can "see" and the sensor can stand being immersed. i.e. Make the sensor face up.
No idea if it's feasible, just throwing it out.

 

Have you experimented with one of those soil moisture detectors? There are lots of DIY projects to build them but you can also just buy one for not a lot of money. It might prove out the concept.

I actually have a soil moisture probe. I'll stick it in some water and see what it does when I draw it out to simulate boiling down.

@keepitsimple: When you say I have to "make a connection to the pot" are you just referring to the effective ground the pot/water provide while the probe is submerged?

 

@keepitsimple: When you say I have to "make a connection to the pot" are you just referring to the effective ground the pot/water provide while the probe is submerged?

Yep. I'm envisioning a totally open pot not Teflon coated where one electrode is the pot or two electrodes where none are the pot.

Knowing what I think I know about mechanical design,the clamp to the edge of the pot would likely need to be contoured to the pot. Just, me; I would not like to mar the pot with a metal clamp.

So, there is a mechanical and electrical design.

 

So I tested my soil probe today. It is a moisture meter made by Rapitest. I took it apart and there is practically nothing inside. There is a needle that goes from 0 to 10, 10 being max moisture. The tip of the probe looks like a rough metal about 1/2" long and then there is a 1/8" wide nylon collar. Above the collar is a chromed rod about 4 inches long, that connects by a wire to the meter. When I put the probe into boiling water, it pegs off the scale. As I draw the probe out of the water slowly, the reading starts to drop to around 3 when it gets near the nylon collar. When nothing but the metal tip is in the water the needle bounces around 0 or 1 or so. In cold water, the meter does not peg out at 10, but it does drop to 0 as soon as that nylon collar comes out of the water.

So it looks like a good test to me. It seems clear that a circuit is being completed by the water bridging the nylon collar. How does this thing work with no battery? Does it have something to do with a bimetallic coil moving the needle?

I would say this test proves the concept as far as detecting liquid level, but how to apply this result to my situation is still a mystery to me.

 

Post a pic of the insides of the thing?

Here's http://rayshobby.net/reverse-engineer-a-cheap-wireless-soil-moisture-sensor/ what I got on a brief search.

 

@keepItSimple: It's the photo of the Rapitest meter that works on the galvanic principal. It's basically a weak battery. I suppose I could post a photo, but there isn't really anything to see. There are two wires going into the box from the probe and those wires go into a small cylindrical hub where the indicating needle is housed. I don't know how they convert the small signal into the needle movement, but aren't we more interested in the voltage generating end of it anyway?

 

I think wayneh put me on to something with the soil moisture meter comment above. As I learned, the meter works on the galvanic cell principle in which dissimilar metals generate a current through an electrolyte. It took me a little while to realize that most every commercial sauce pan/ stock pot is made of aluminum. I had a section of copper tubing so I measured the voltage generated between the aluminum and copper in boiling tap water! This came as a revelation to me because when I first came here I posted a circuit that forced a current down the probe to simulate ground. By using the aluminum pot and copper rod (both food safe materials) I can generate the charge without any battery other than what is required by the alarm (and maybe the circuit??).

I boiled some water and found a very steady reading of about 1100 mV. It didn't matter how deep or shallow the copper tube was in the water, it remained at 1100 mV. Of course when I pulled it out of the water simulating boil off, the reading was 0 mV. I tried to measure the milliamps, but with my probe held against the side of the pot the reading wall all over the place.

So can someone help me figure out how to turn this voltage into an alarm when the voltage hits 0? I have some good ideas from earlier in the thread but I think I need a rework of how to do this based on the dissimilar metal concept.

 

I tend to use comparator circuits a lot, such as the LM339. That has 4 comparators in one IC package. (You could set up 4 different levels with that. ) Take a look at the LM3914 also. That's what you'd use as a bar graph style voltmeter.

You would set one input at a reference of 500mV and connect your "battery" to the other input. The state of the comparator would change as the battery voltage falls on one side or the other of that reference.

You should check what reading you get with stainless steel, too. That's very common in a pot.

A comparator has a very high input impedance, similar to your voltmeter. Very little current is needed to drive its inputs.

Copper isn't completely inert in cooking applications. You may want to research where/why copper is either used or not used. It's commonly used for candy, and in brewing it has a purpose as well.

 

This is coming a little late to this thread, but I've had success with using cheap ultrasonic sensors (HC-SR04) from ebay as water level sensors, the advantage being that the "sensor" doesn't need to be in contact with the measured fluid. One situation was corrosive environment (salt water swimming pool) and the other was a flowing stream with lots of surface debris. Disadvantage I suppose is requires a microcontroller, just a simple and cheap PIC is all that is necessary, but still maybe doesn't suit your situation. If your interested, I can post link to circuit, code and an example.

 

I boiled some water and found a very steady reading of about 1100 mV. It didn't matter how deep or shallow the copper tube was in the water, it remained at 1100 mV.

Today I got my grandson to do a similar experiment. For a probe in the form of a ~3mm diam brass rod in slightly salty room-temperature water in an aluminium pot we got about 0.55V, with the brass being positive with respect to the pot. Using a stainless steel pot instead, we got about 0.01V, but less steady, the brass again being positive with respect to the pot.
On this basis a simple comparator circuit should be able to give a reliable indication of whether or not the probe is in a (slightly) conducting liquid in a typical saucepan.

 

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.

You might try using a gun-bluing kit on your probes. There are 2 kinds of rust-- iron oxide, and nitrogen-oxide. Neither can rust the other. A gun bluing kit essentially rusts a gun nitgrogen-wise, so it can no longer be rusted by ferrous/oxygen rusting. Plus it will give your probes a neat color. Test it on an inexpensive probe first.

 

You could try this http://www.linear.com/product/LT1789 IA and a CR2032 battery.
You;ll need a gain resistor, bypass caps and likely two 10 M resistors from the - input to ground and + input to ground as well.

You may end up with corrosion and copper oxide may have a higher potential than clean copper.

Now, we have to find a micro-power comparitor..

 

Thanks for all the responses! I wanted to throw out some follow up measurements I made today. I think my results were similar to Alex (thank you for the confirmation, Alex!).

Using a 1/4" copper tube in about 4" of cold water in an aluminum pot, I got a reading of around 400 mV, I guess 0.4 V. This bumped around a bit because I realized that since I was holding the probe to the pot with one hand, and the other probe to the copper with the other hand, I was getting a reading across my body. I used electrical tape to secure everything the best I could. I think the reading was affected by the temperature of the copper tube. Anyway, 0.4 V was the reading with electrical tape. Next, I boiled the water and took another reading. It came out about 1.0 V, so clearly it works "better" with hot water.

Doing the same thing with a stainless tube of the same size, I got about 0.0 V with cold water and 0.4 V with boiling water. I agree that using stainless would be better than copper, so maybe I just have to stipulate that the probe be inserted into the liquid after it has started heating up first.

Given the above, I need to look up everything suggested to me and try to understand better. Please post any additional info you would like. I'm sure I'll be back soon with more questions!

P.S. I would like to repeat that I'm trying to do this with only a small battery, and not a/c power.

 

here's (Some variant of the LT6700) http://cds.linear.com/docs/en/datasheet/6700123fh.pdf an idea for a comparitor.

These IC packages are tiny. www.proto-advantage.com can solder parts available from digi-key onto standard SIP spacing PCB's if they have the right adapter. They procure the parts as well.

So, at least you can get a LED varient to work. e.g. LED on or OFF when out of pot.

With a little xtra work, you can use wires (suggest PTFE insulated) to say a 9V battery type of alarm.

The design using the IA and comparitor isn't done. It just satisfies, a single supply, micro-power comparitor and LED.

Issues are low power means low voltage and usually 1.8 V is the absolute lowest one can go. 3V is a common supply from 2 AA, 2 AAA and the coin cells. The low power stuff usually doesn't like to exceed 5V in supply voltage.

Low power can be a challenging design.

 

You might try using a gun-bluing kit on your probes. There are 2 kinds of rust-- iron oxide, and nitrogen-oxide. Neither can rust the other. A gun bluing kit essentially rusts a gun nitgrogen-wise, so it can no longer be rusted by ferrous/oxygen rusting. Plus it will give your probes a neat color. Test it on an inexpensive probe first.

Maybe something to look at once I have a prototype working. Seems it might be difficult to do if I use a stainless probe.

 

Now, we have to find a micro-power comparitor.

Not sure if that's really necessary. I was thinking along those lines, hoping to minimise standby current and thus avoid the need for an on/off switch, simplifying making the whole caboodle waterproof. However, the default state of the unit is out-of-liquid when not in use; which implies the default buzzer/alarm state is 'on' and guzzling most of the current. Hence a switch is mandatory, so minimising current is less important. A LM339 should be fine (at 0.8mA current draw) IMO. But if you really want a low-current comparator, how about a LMC7211A1 ?