LDC0851 Liquid Level Designs

Hi Na Na,
I will forward this post to my colleague who has worked on this application; he will get back to you shortly.

Hi Luke

Thank you for your advice.

We made a prototype like your figure, and it worked.

But we would need information for the "principle" of this configuration.

Why doesn't it need inductance?
How does the core work inside the device?

We would like to use LDC0851 more properly.


BestRegards

Hello Na Na,
Glad to hear you got your prototype working! The fundamental principles are the same as using the LDC0851 normally. The output will switch LOW when fsense > fref. Similarly when fsense < fref the output will switch HIGH. Increasing the capacitance on the reference node effectively slows the frequency down until fsense > fref, then the output will switch LOW. You can tune the frequency with the choice of resistors and capacitor. Hope this helps clear things up.
Regards,
Luke

Hi Luke

Thank you for your comment.

According to your schematic, I image the attached rough sketch for LDC0851 liquid level sensing using CR oscillation.

sketch.xlsx

Is this understanding correct?

(Sorry, Lsense and Lref are reversed)

BestRegards

Hello NaNa,

You are correct.

Regards,
Luke

Luke,

Thank you very much for this simple circuit diagram.  If I want to sense water inside a vertical rubber or plastic hose or tube, I should be able to attach a lead wire to a metal band that wraps all the way around the hose.  If this hose is coming from the bottom of the tank and going to the top of the tank, it will be much like a sight tube.  Since the LDC0851 works on differential from the reference, I can see how it would work well when the water level drops below the upper metal band or plate.  That's fine once the system is established.  What happens if the tank is new to a customer and there is no water in it.  The manufacturer of the product does not want the product to operate unless there is a minimum fill of water in the tank.  If the Reference plate is just as dry as the Sensor plate because there is zero water, won't the LDC be fooled into thinking there is nothing wrong?  Is there a way to overcome this without adding other kinds of sensors?  A bit more circuitry is fine, but there is no opportunity for a float or any kind of contact with the water.  The water may also be of varying quality.  Some will come directly from low quality ground wells and is non-potable.  Some will contain high levels of calcium and lime and other dissolved minerals or chemicals.  Does this present any sort of problem?

Thank you for your help.

Ken

Hi Ken,
You can size the resistor values so that with no liquid present, the output is off (potentiometer is easiest for prototyping). This is akin to having a slight frequency offset which guarantees the part will be in a certain state in nominal conditions. Alternatively, you can use the ADJ pin which effectively increases the frequency of the reference channel. See section 8.3.2 Threshold Adjust Mode of the LDC0851 datasheet for more info on the ADJ pin. Note that the switching point occurs when FSENSE / FREF > 1. So increasing the ADJ value will increase the reference frequency putting the LDC0851 in a known state with no liquid present.
It's also important to note that the frequency decreases with liquid present. Therefore, to further enhance the system you can size the FREF metal sensor to be larger than the FSENSE metal plate. This will result in a larger frequency shift for the reference channel as the liquid passes over the metal sensors which will help guarantee the FS / FR > 1 condition.
Regards,
Luke

Thanks Luke,

This is sounding much more like what I am looking for.  I believe I can get a breadboard circuit made up now.  I need to locate a chip carrier for a WSON-8 that is breadboard friendly.  I am not sure about all the capacitors shown in the circuit.  Is it safe to start with nothing?  Do you think I will need any?  If yes, what are your predicted values for which locations?  Are there specific types of caps that are preferred?

There will be a substantial amount of stainless steel within one to two inches of the hose I want to wrap with the metal band as sensors.  Do you think the nearby metal will affect the operation of the circuit?

Thanks so much for all of your input.

Hi Ken,
I would definitely leave a footprint for the sensor capacitor. It is helpful to have a discrete capacitor (~30pF) for stability, NP0/C0G type is best. I wrote optional because the more capacitance you add, the less sensitive the circuit will be. And note that the other capacitances drawn in the circuit (board parasitic, pin package parasitic, metal plate capacitance) are theoretical, therefore they do not need additional footprints. You can place the footprint for the sensor capacitor next to the IC. Refer to page 29 of the LDC0851 datasheet for example device/component layout.
The additional stainless steel could impact the sensitivity if it is too close to the sensor. 2-3 inches is likely ok for small sensor electrodes, but it is always best to prototype to see the response.
Regards,
Luke