Passing 2 wires into a hot water cylinder

[dannydc]

Hi everyone,

I'm trying to find a robust way of passing 2 wires into a copper hot water cylinder so I can put some DS18B20s in there to collect temperature data. The sensors can be daisy-chained and powered from the onewire line if need be, so 2 pins is enough.

Does anyone know of a good way of getting those pins through some kind of compact bulk-head connector that I can install into the immersion heater cap? It needs to be able to withstand 3.5 bar (cylinder max working pressure) and 60°C ish.

Thanks,
Danny

[Benta]

Why do the wires need to into your vessel?
Make a water/pressure proof probe that you can stick into it. Place the sensor inside that.

[themadhippy]

I can install into the immersion heater cap?

Make a water/pressure proof probe that you can stick into it. Place the sensor inside that.

If theres an immersion heater theres a 99% chance theres already a water tight tube going into the cylinder for the existing  stat

[ConKbot]

A "thermowell" would be the usual method of getting a temp probe into a vessel https://www.omega.com/en-us/accessories/protection/c/thermowells just as examples, since most are good for a hideous amount of pressure, and they charge accordingly.

Im not familiar with your setup, but if you can find a thermowell that fits your application, and drill/tap your heater cap, you can have your sensors at ambient pressure, so no need to worry about connector ratings other than temperature.

[wraper]

Nobody in sane mind passes wires into pressured vessel. For temperature sensing usually it's something like this https://www.aliexpress.com/item/1005005676221430.html

[antenna]

I've soldered to spark plugs before...

[Geoff-AU]

I would question the need to sample from the pressurised wet side.  How accurate do you need it?  I have a hot water tank with 3 sensors on it, 2 in thermowells and 1 is jammed between the insulation and the pressure vessel. The thermowells only read at most 2.5 degrees higher than the surface temperature.

[max_torque]

I have a basic HW tank without more than one temp sensor positition, so i simply made a "probe" that passes in through a 22-22-15mm threaded(olive sealing) unequal T fitting and consists of a short 15mm to 8mm adaptor, that then has a 1m length of 8mm dia copper tube sodlered to it and that tube has it's bottom end capped with a soldered on 8mm cap.  I simply push a chain of temp sensors PT100's in my case, int othis tube, and the entire assy is then fitted onto the top vertical hot water outlet pipe, and the long 8mm tube then just hangs down inside the cylinder.  Because the vetrical tube is only 8mm dia it does not excessively impede the normal flow of hot water out of the 22mm exit tube


This means you are using fully rated parts and the actual sensors are dry. The copper tube is more than conductive enough when surrounded by the high thermal mass / low thermal impedance water so the tube effectively maintains the temp gradient of the water inside the cylinder.  I have 5 sensors in this tube and from that i can calculate the total energy stored in the cylinder with some basic maths and averages :-)  You can also easily access the sensors should you need to by just unscrewing and lifting out the sensor tube from the top of the T

      |  |
    | |  | |
    | |  | |
   |  |  |  --------
   |  |  |        ------>
   |  |  |  --------
   |  |  |   |
   |  |  |   |
      |  |
   ^ |  | ^
      |  |
      |  |
      |  |
      |  |
      |  |
      |  |

[tmd63]

My HW Tank has one sensor strapped to the outside of the tank! Copper is a very good conductor of heat so the sensor is directly on the tank under the insulation jacket.

[Gyro]

Likewise, this is the standard cylinder stat for an external heat source, typically a gas boiler. Although the Copper is highly thermally conductive, it is also very thin, so there is no problem detecting the water thermal gradient within the cylinder, ie. colder at the bottom and hotter at the top, from the outside surface.

[pcprogrammer]

I'm trying to find a robust way of passing 2 wires into a copper hot water cylinder so I can put some DS18B20s in there to collect temperature data. The sensors can be daisy-chained and powered from the onewire line if need be, so 2 pins is enough.

Does anyone know of a good way of getting those pins through some kind of compact bulk-head connector that I can install into the immersion heater cap? It needs to be able to withstand 3.5 bar (cylinder max working pressure) and 60°C ish.

The DS18B20s is not the most precise sensor so the measurement will already deviate from what the temperature really is, so no need to immerse the sensor into the water. Like others wrote, just stick it on the outside or use a fore mentioned thermowell.

Another issue with immersing a sensor like this is corrosion or deterioration of the wires due to the hot water. I have experienced this with the sealed and water proof sensors in my central heating system where I used them to measure the temperature in non pressurized storage tanks. Heated up to 85 degrees the wires did not last long and became brittle and the core eroded and became non conductive. Changed the setup to have the sensors in sealed pipes that are immersed and keep the sensor and leads dry. Also makes it easier to replace them in case of a defect.

[2strokeforever]

Find something like this
https://www.ebay.ca/itm/285509878582

its just a tube with the sensor in it, and a compression fitting that seals it up. they are very proven and not much to go wrong.

[tom66]

Why not just put a temperature sensor on the hot water output of the cylinder. It will be very close to the actual temperature of the cylinder water, since there will be some conduction even without active hot water flow.  If you insulate that area well, it will be even more accurate. 

[ejeffrey]

Thermowells have a number of big advantages that make them the overwhelming choice.  They precent the hot liquid from attacking the sensor, wires or insulation.  They prevent the sensor from contaminating the liquid (important if it's food!) and avoid nooks and crannies that can trap liquid and make it difficult to clean.  It's much easier  to remove and replace broken sensors or just take them out to calibrate. There are enough advantages that it's the most common way to do this and thus the easiest.

If you really need a feed through, swagelock fittings can be used for this.  Just use a solid rod the right diameter for the ferrule and connect to it on both sides.

[max_torque]

When you measure the outlet (top) water temperature of a typical domestic hot water cylinder it's not a lot of use from a control perspective because it pretty much reads the same value until the tank is completely empty of hot water and then it suddenly reads close to the inlet water temperature!

This is because the water experiences a density change as its temperature changes, and the water at the top is always the hottest water, the incomming cold water simply sits below it.  Slowly over time, you find that top temp will fall as it looses heat to the colder water below it, but this isn't really enough to put in place much of a decent control system.

For example, my system pretty much sits at a top (outlet) water temp of between 55 and 50 degC, yet the 1/3rd point (1/3rd of the distance from the bottom) and the 2/3rds point experience much greater swings, and you can use them to estimate how much water is in the tank of any given temperature, ie the remaining capacity of the tank, and control heating as necessary.



The lower temperatures clearly show when hot water is drawn from the system, as the "slug" of cold water comes in, immediately drops the temp on the lower sensors, and then slowly warms up over the next few hours as the top temp drops very very slightly.  During heating, the lower temp jumps right up, by more than 20 degC, yet the top temp onyl changes a couple of degrees

[zapta]

Have you consider thermowells  They come in many shapes, with or without threads, are hollow, and allow to insert temperature sensors from the outside, keeping the sensor dry.

https://duckduckgo.com/?q=thermowell&iax=images&ia=images

[Haenk]

Measuring in pressurized environment is done in our heating environment - using "Fernheizung" (translation?). I.e. the building is heated with pressurized hot water, distributed in the whole town by pipes and a central heating facility. For small homes, sensors are strapped to in/out pipes. Larger facilities do have sensors INSIDE the pipes, so these should be readily available, at least in Germany.

[tooki]

Measuring in pressurized environment is done in our heating environment - using "Fernheizung" (translation?). I.e. the building is heated with pressurized hot water, distributed in the whole town by pipes and a central heating facility.

FYI, that’s called “district heating”. (Translates to “Bezirksheizung” literally.)