Sensor for a Water Heating Vessel

[Little Bill]

I was working in a place doing an addition and got asked to help with a problem.
They have a large water heating vessel that is remote controlled. Basically it's just a temperature controller, relay, and contactor.
The heating element burnt out and they ordered a replacement and wanted me to hook it up. After getting it hooked up nothing would work past the temperature controller.

I didn't have any schematics so I just tried to figure it out. Turns out there is a level sensor in the tank/vessel. It has a low voltage wire connected to the sensor and ran back to the relay. I had continuity on the wire and thought it was good. The sensor is just a probe that goes in the side of the tank. It is supposed to keep power from reaching the heating element if the water level is too low. There is no voltage in this sensor or wire, just a connection to the control.
I finally talked with someone from the mfg. and he had me try a few things.

Turns out that along with the sensor wire back to the controller, the tank should also have a bond wire back to the ground bar in the control cabinet. I'm trying to figure out how this works. The only thing I can come up with is the controller is reading resistance. The low voltage wire, along with the bond wire, makes a loop and resistance is read through that loop. If the probe is in the water the resistance would be different than when the water is below the probe. The controller senses this and either allows the relay to pull in or not, depending on the reading.
Am I even close on how this works?

Also, there is a temperature sensor. I think it either a thermocouple or RTD.
It also controls whether the element receives power or not but is separate from the level sensor. Just thought I needed to add this to the description.

 

[don_resqcapt19]

Since the equipment was working in the past, I would not expect that you would have to add a bonding jumper to make it work now. What else has changed?

 

[Little Bill]

Since the equipment was working in the past, I would not expect that you would have to add a bonding jumper to make it work now. What else has changed?

All that I know that changed/happened was the element burned out and was replaced.
The factory tech had me do a "temp fix" by running a jumper from the relay terminal that the sensor cable is on to the grounding block in the control box. That allowed everything to work.
He told me that there was supposed to be a bonding jumper when it was set up. Now how it was working before without it, I don't know. The owners of the equipment did tell me that they often have trouble with the sensor signal not allowing the heater to operate.

I just left the "temp fix" for now until I can run a jumper from the tank to the control box grounding block.

I'm still wondering how this sensor process works. Is it resistance like I mentioned in my OP?

 

[don_resqcapt19]

I would expect capacitance but not sure.

 

[hbiss]

The sensor is just a probe that goes in the side of the tank.

Low water cut off. All commercial boilers and many residential boilers are required to have them by code. All work similarly. In your case I would remove the sensor probe from the tank and see if it is coated with any kind of corrosion or crud. Since you were able to get the controller to work by jumpering the probe terminal to ground that tells us they are using the resistance to ground to sense when the probe is submerged. Do check to see that there is continuity between the grounding block terminal and the tank where the probe screws in. If all is good do like I say and pull the probe and clean or replace it.

-Hal

 

[gar]

160730-1512 EDT

don:

I would really expect conductive rather than capacitive.


Little Baill:

If we assume that the "control" has some connection to earth ground or EGC, then it should be possibe to quickly determine whether the control works by shorting the "low voltage wire" to an EGC location, possibly the "control" enclosure..

If the "control" functions, then a return path (ground, EGC) needs to exist from the tank.

If the "control" does not function, then further troubleshooting is required.

I would disconnect the "low voltage wire" from the "control", then measure for either or both AC or DC voltage from the "low voltage" terminal to the "control" enclosure, and a separate measurement to an EGC point. If at least a few volts are measured with a high impedance meter, then I might question whether the temperature sensor is functioning.

Why do I think it is a conductive sensor instead of capactive? Most water is moderately conductive, even distilled water exposed to air is slightly conductive. Conductive sensors are easier and less expensive to build than capacitive. Unless a capacitive sensor is placed directly at the probe or only a short distance away and connected with a coax cable the shunt capacitance and variability of that capacitance can be a major problem in setting a threshold level.

Further at low frequencies, 60 Hz, the capacitive reactance of the probe when in the water would be totally swamped by the much lower shunt resistive component of the water. So depending upon how the capacitive sensor works it still would end up being a conductive sensor.

Much higher frequencies, possibly 100 MHz or higher, would be required to make the capacitive component significant relative to the resisitive component of the immersed probe. At these high frequencies the stray capacitance of the cable and its variability are going to be major problems.


hbiss:

Where did Little Bill ever say he made the circuit work by shorting the "low voltage" wire to ground. He did ask if his theory on it being a conductive sensor was correct.

Your response was good and a probe that developed a high resistance on its surface could be a problem.

As a quick check I might try a resistance measurement of the probe to the tank with a Simpson 260 on the Rx10,000 range. Obviously with the wire to the control disconnected. My meter on this range is a source voltage of 6 V, newer ones are likely 9 V. With my city tap water I read about 15,000 ohms between two #20 wires spaced about 1" apart and immersed about 1" into the water. With DC, which the Simpson is, the resistance reading will increase with time. Change the polarity and it goes back down. With AC you do not see this time effect.

.

 

[hbiss]

Where did Little Bill ever say he made the circuit work by shorting the "low voltage" wire to ground.

In his second post:

The factory tech had me do a "temp fix" by running a jumper from the relay terminal that the sensor cable is on to the grounding block in the control box. That allowed everything to work.

-Hal

 

[don_resqcapt19]

160730-1512 EDT
don:
I would really expect conductive rather than capacitive. ...
.

The boiler low water level devices that I have worked on have either been capacitive or conductive, but the conductive ones operated at about 500 volts and required two wires. The capacitive ones used either a single wire or a coax cable.

 

[gar]

160730-2504 EDT

don:

An original Warrick type conductive sensor consisted of an isolation transformer that might have a 500 V secondary, a moderately sensitive AC relay, and a probe. Part of the circuit was usually a "ground" path. Thus, some might call it a single wire probe.

By using a high voltage it was possible to get good operation with ordinary tap water and switch a fractional HP motor.

About 1960 I developed an electronic circuit that could control the same relay, work with a 10 V probe voltage, and have a threshold sensitivity as high as 1 megohm, if desired. Further it was moderately insensitive to input line variation from 95 to 135 V. The probe voltage was AC in contrast to DC and therefore did not have the probe polarization problem of other competitive low voltage probe circuits that used DC on the probe.

Also did capacitive sensors. This was more for non-conductive liquids or for proportional level monitoring. If a conductive application had a major probe contamination problem, then a capacitive sensor might be useful.

The Warrick brothers sold the company somewhere in the 70s I believe. Their father was the founder of the company, and when I consulted with them in the 60s the company was located in Berkley, MI on 11 Mile Rd.

.

 

[gar]

160731-2349 eDT

hbiss:

Thanks. I guess my reading of that line did not make me think clearly about what was said.

I want to know what the voltage reading is on that relay terminal relative to the ground bus when the wire to the probe is disconnected from the relay coil terminal? Both the DC and AC voltages are needed. Any DC component should be stripped from the AC voltage measurement. A Fluke 27 or 87 does this. A Simpson in AC does not, but it does when Output is selected.

What resistance value, when connected from the relay coil terminal to the ground bus, just causes the relay to pull-in?

.