Yep, nothing there "on the surface". That's part of the problem. The only part that's visible is under "packages" because that's where I've been playing.

Eventually, you can find "My parts" with two of them defined: "LT6700CS6-1#PBF copy" and a "IC_REF2030 (SOT-23-5) incomplete". They are hard to find and/make and that's what I find annoying so far. Not intuitive at all. So, it's the comparator and the reference.

I just modified the REF part. The two other ones needed are the optoisolator and the OP Amp. This is the steep part of the learning curve - creating the packages and symbols. Fortunately, the package outlines are more easily copied.

It's like, there is no point building a brick house without bricks or without the selected windows.

 

Yes, I saw those but didn't recognize any of the numbers from our discussion. Anyway, I figured you were just getting started on it. Glad at least I could give you a project that will help you learn their software!

 

The LT6700 is the series of the comparator. The one we talked about in the very beginning. The real part number can be a mile long. Then there could be the digi-key stock number. The variations are LT6700-1, -2 and -3 and then there are package variations. An entirely different part number is used for the single version. For the OP amp, the suffix gets weirder. When I used it, I could not find anything on DigiKey.

One of the early things I did was to wire some parts together/

 

...... in the meantime, here's a simple circuit that might be worth try. A supply decoupling capacitor might be needed but isn't shown. A and B are the connection points for the stainless steel probe and aluminium pot respectively (assuming the electro-potential of the probe is more positive than that of aluminium). Standby current is about 80uA, so a PP9 battery should last around 6 months on standby, making an on/off switch arguably unnecessary (which would simplify sealing a circuit enclosure).
Trim1 sets the general threshold voltage at which comparator U1a trips. Trim2 sets the amount of hysteresis. The MOSFET type will depend on the buzzer current, and a snubber or catching diode across the buzzer may be needed, depending on the type of buzzer.


Edit: Oops, the switching logic is wrong. See revised circuit in next post.

 

Right. Here's the revised circuit, with the logic corrected and the addition of input protection, output protection and supply decoupling. An on/off switch is now included, to silence the buzzer when the circuit is not in use. Standby current (circuit in use but buzzer not sounding) is ~30uA.

 

Alec: Thanks for taking the time to put that together! Regrettably as a Newbie I'm not able to comment much other than to hope that this circuit is something similar to what KeepIt is putting together.

 

I look forward to seeing the KISS solution.

 

Thanks Alec.

I played with Easyeda again last night for a couple of hours and made an attempt to create a library symbol for the Instrumentation AMP. I don't 100% like how it looks, but it's a first time. The pin name and number have a fixed layout, so they sit on top of lines.

I also tried to make the optocoupler, but the symbol that existed was a "Spice symbol" and it won't copy to a librray symbol.

 

Alec:

See: http://www.analog.com/media/en/technical-documentation/data-sheets/AD8231.pdf#page=23

Can you come up with good values for C1, C2, R1 and R2; Note that R4=10K and R3=5.9K. Don't plan to use R5 or C3 or C4.
R6, R7 would be 200 ohms minimum.

Any frequency somewhat below line frequency would be OK, If I can't come up with anything, I'll use 32 Hz. 3.2 Hz might be obtainable with real components. The object is to amplify DC only and particularly reject line frequency of 50/60 Hz. It should also be able to reject "bubbles".

 

Got R1,R2 = 158K and C1, C2 = 100 nF for a 10 Hz -3 db frequency. Response is down about -30 db at 60 Hz, so the gain is down by a factor of about 31.

 

The comparator resistances:

3-1.5+0.8 for greater than 2.3 V to > 400 mV; Resistors are 4.12 Meg ohms (to ground) 1% and 866K to +3 V
3-1.5-0.8, for < than 0.7 V to < 400 mV ; 2.87 Meg ohms 1% to ground and 2.15 Meg ohms to +3 V

Basically want the sum to be pretty high. About 5 Megs seems reasonable. They are 1% resistors. Did not check availability.

 

Can you come up with good values for C1, C2, R1 and R2

Your post #190 values look good.

 

Worked on the CPC2006 OPTOMOS relay symbol tonight. Again, not the greatest, but OK. No footprint. It went a little faster.

 

The circuit is starting to take some shape at EASYEDA. You can take a look there. You have to remember to "snap to grid" and change the grid if you have to.

Wire tools and drawing tools are different. The program gets "stubborn" once in a while and won't do what it's supposed to do. Saving and re-opening usually fixes it.

The day I had trouble separating the pin name/number worked another day, so it may just be one of those glitches. Replacing a part with a modified (edited one) didn't go simple. Again, this is where refreshing the browser works. e.g. You can't get the arrow cursor back.

So, I had to clean up the AD8231 package to get the wires to wore nicely. So, schematic consists of the IA and input circuit and Reference.

I need to clean up the reference package and set the gain and enable signals and then add the filter.

 

Well, in the midst of lots of home problems, I managed to use easyEDA to put together the MEAT of my version. This needs to operate at 5 to 5.5 V or standard values for 5V logic because of the REF IC.

Another "schematic" is intended to make it operate off a 9V battery (probably from 5 to 20 Vdc). The buzzer may reduce the upper end of operation.

The "alarm section" should be pretty straightforward. I don't plan to use a FET to drive the buzzer, although the option is open.

Alec: I noticed you put a reverse biased diode on the buzzer. What's your thought on this?

I think the PCB should accommodate an optional isolated contact closure and I may attempt to include one. Uses could be for systems that say send an SMS on an alarm.
If the design is workable, I may add it.

This "main section" uses a low power instrumentation amplifier which amplifies (in this case buffers =, x1 gain) the difference of two signals.
This offers a number of advantages:
1) Canceling of common mode noise (50/60 Hz);
2) Two or more "power supply shared" systems could operate together. (the connected burner grates would not cause ground loops). With long wires, pickup is extremely likely. Twisted wires would help too. This, just remains to be seen.
3) Plus and minus polarities would automatically be accounted for. It's easy, so why not. probe polarity then would not matter.

The design is not suitable for induction ranges. Aluminum won't work in an induction range. There is NO RF filter in the front end. The pot must be Aluminum for the highest sensitivity and a threshold of 800 mV.

The IA has pin selectable gains and an uncommited OP amp. The OP amp allows the implementaton of a filter, so that the "conditioned" signal is nearly DC (10 Hz bandwidth). This will further cut 50/60 Hz interference.

The basic problem with electronics is that ZERO isn't obtainable, 1e-12 might be, but not zero. PC traces have inductance etc. Identical components aren't obtainable either. IC's have made "identical" pretty close. Temperature is close too. Components like resistors can be "laser trimmed" after they are made. Some circuits like the IA require matched resistors for best performance.

So, bypass caps prevent oscillation and compensate for the inductance of the PC traces and do the equivalent of eliminating "hesitation" in the carburated internal combustion engine. The 200 ohm resistors are for protection for an unpowered device and the 10 M resistors from each input to ground, give a place for the very low input bias current to go.

This piece is buildable on it's own and needs a 5V supply. The 5V supply is reduced to 3 and 1/2 of 3 Volts for the IA. The output is offset to 1/2 of 3V or 1.5V with the leads disconnected. So, the output relative to ground is 1.5V + V(probe); If the probes are reversed, then V(probe) could be negative.

The comparitor thresholds in the other circuit are then greater than 1.5+0.8 V and less than 1.5-0.8 V

Things may slow down for a bit while I attend to some personal matters. It did help me relax.

Note:
In EasyEDA, I just chose 1206, it's a physical size, SMT (surface mount) discrete components because they are large. I also cannot guarantee the availability of such components either. The PCB layout pads may or may not be correct. Primarily, my intension was to render a schematic.

My learning curve primarily involved:

1) Spice and schematic symbols are very different. You can't use someone else's spice symbol for a schematic symbol even if if it looks like exactly what you want.

2) You MUST use snap to grid when drawing anything,

3) "My components" that are IC's begin with IC. Listing "My components" doesn't really seem to work.

4) sometime EASYEDA just doesn;t do what it's supposed to do. You have to save and refresh the browser. Known bug, apparenty.

Ideally, you need to "layout" a really simple circuit from scratch to really understand the process. Let's say it's a coin cell, led, pushbutton and resistor. You can first use their parts and do it again making your own parts.

Conceptually, I know what's involved, but I have only used the old "tape" method that I learned at HP/Agilent/Keysight in a crash course and I layed out 8 or so small boards for a work project using an old DOS program named Easytrax. I did some ground plane stuff by modifying the generated postscript file which is pretty hard core. I did some "playing" with DIPTRACE. Eagle gave me a headache. KiCAD is keystroke based and I don't memorize well.

I have Target 3001 http://www.ibfriedrich.com/en/index.html , but don't have a system to run it on. I played with it under emulation for a few hours.

My future endeavors involve really weird shaped PCB's and reverse engineering, thus Target,

A CAD program is in the cards too, but the one I really like (Vectorworks) runs on Windows only and is too expensive for home use. I didn;t have enough time to evaluate BricsCAD, but it seemed likeable. I'd almost like to run a license manager at home as well, but that's another problem.

 

Alec: I noticed you put a reverse biased diode on the buzzer. What's your thought on this?

Well, there are buzzers and there are buzzers. An electromagnetic type might have considerable inductance, whereas a piezoelectric one wouldn't. Better safe than sorry.

 

Well, after some additional time, here is my "Alarm module"

If you wanted to de-bug it stand alone, it would operate on 9V. i.e. the 5V supply can be 9V for testing.

You would need an OP-AMP that's unity gain stable or an adjustable power supply fro 0 to 3V. An AD8663 http://www.analog.com/media/en/technical-documentation/data-sheets/AD8663_8667_8669.pdf would work, but it's likely overkill. Unity gain means the - input is tied to the output and the input is applied to the +input of the OP amp. The input could be the wiper of a 1 M 10 turn variable potentiometer that sits across a CR2032 battery with a switch. This would give you a 0-3V low impedance reference to check out the circuit design.

Alternatively, we can use the 5V supply and divide it down in the same way for a 0-3V source or even a 0-5V source for testing.

The 0-3V signal would be supplied to C(sig) out to test the alarm at > 1.5+.8 and < 1.5-.8V volts or >400 mV and <400 mV at the comparator.
the comparitor will withstand more than 9V input if unpowered.

I DID NOT rigorously check the comparator design. It's output might be inverted. I tend to get confused with comparators anyway and the datasheet isn't much help. if it's inverted just switch the +- inputs. This particular comparator doesn't need a low Z input, but the input to the dividers do.

The two resistor dividers can be reduced to three resistors. 1% resistors are used here. They could be replaced with 10 turn trimmers to for adjustable setpoints.

Again, a bypass cap and a diode for buzzer protection (thanks alec).

I included another contact closure which as ~10 ohm resistance which can be eliminated if the resistor is changed. The second contact closure can possibly be used for other things. One possible idea, is to connect it to a device that sends an SMS message, Interface to another alarm, voice module or LED. probably overkill.

Using the topology, it would be very easy to add a LED turned on from the second contact and an Alarm silence switch. "Alarm silence" would prevent the buzzer from sounding, but the LED would light. The aux contact function would be removed or another one could be easily added.

The buzzer has to have a current rating less than the optomos part. FETs for switching is another possibility.

If the buzzer operates at 5V, the system, as presented so far, could operate from a 5 V wall wart or a 5V wall wart and a 9V battery.

The design is pretty straight-forward. The resistor dividers should be double-checked and the polarity of the comparator inputs needs some checking. I have a 50% chance of getting it right.

The easyEDA design is not ready for prime time. The footprints are not guaranteed to be right. The part numbers aren't.

Even as it stands right now:
The buzzer footprint isn't right. It's a chassis mount buzzer which could require a connector and a grommet as well as screws. It MIGHT be able to protrude from a hole in the case.

The aux contact is an option. I'll leave it at that.

For a single-channel design, one might have an (on)-OFF-on switch and a ON-OFF (buzzer silence) per channel. The momentary (ON) could be a battery test or just create a LOW BATTERY LED.

I like to throw out all ideas, no matter how bad they seem and then start removing or incorporating a variation of them.

The "production board" could just leave out the OPTOMOS part, and connectors and that option might be installable by moving a jumper (SMD resistor), removing one and installing the optomos part and header and connector. Cutting a trace is an option and so are whats called a "split pad" or both. The PCB and case could just accommodate, but never use the option until someone asks.

Your not there yet.

 

Your not there yet.

Well, umm, since you lost me right after "here is my "'Alarm module,'" I'm going to trust your judgment on that one!

 

I'm basically going to leave the design "as is" until I get some "direction" as to what's needed.

What to use to drop the 9V battery to 5 V needs a LOT more thought. For debugging only, an L78M05 https://media.digikey.com/pdf/Data Sheets/ST Microelectronics PDFS/L78MxxAB,AC .pdf could be used, but it's a really stupid choice because it may only operate to 7V input.

One choice I looked at was made on a 4 layer board and the evaluation kit was nearly $175.00. They are finicky and you might have to worry about RFI emissions.

A 5V supply (USB charger) from a cell phone charger should be all you need or something like this: http://www.mouser.com/ProductDetail/Mean-Well/GS06U-1P1J

ideally, the best thing to do is to find an SMT thru-hole adapter and have www.proto-advantage.com procure the IC parts from digikey and mount them for you.

You can use stuff like this https://www.digikey.com/product-detail/en/capital-advanced-technologies/6109/6109CA-ND/151945 to say mount the 200, and 10 M resistors and plug it into a socket on a thru-hole breadboard. You can use a solderless breadboard e.g. where https://www.digikey.com/product-detail/en/bud-industries/BB-32621/377-2094-ND/4156445 where you just plug stuff in and wire with solid wire.

This https://www.digikey.com/product-detail/en/vector-electronics/45P80-2/V2022-ND/1886444 has been my favorite for a long time. Lots f different varieties exist. Wire wrap wire can be used to make connections or the square pads can be bridged.

 

OK, I see the options for construction. I understand using a plug in power supply for testing purposes but this presumes you already have a battery in mind that will work when it comes time to replace the power supply?