casting / potting electronics to waterproof a soil water sensor

Thread Starter

micromet

Joined Sep 10, 2016
9
Many soil water sensors use a casting resin or epoxy to cover the electronics on the sensor head (great example from Pino-Tech shown below).
https://pino-tech.eu/product/soilwatch-10/

Does anyone have some information or link on how to cast your own protective sensor head (mold construction, epoxies/ resins, mold release, etc.) etc). I would really like to learn a DIY version of this approach.

Thanks,

 

ebp

Joined Feb 8, 2018
2,332
I assume you are referring to the black end.

That has probably been done with injection molding with a thermoplastic, most likely plasticized polyvinyl chloride (PVC). For high durability cables, thermoplastic polyurethane is popular. Epoxy would not be suitable if the strain relief for the cable is integral with the molded part.

Commercial molds for injection molding for low to moderate volume production are typically made from aluminum. These days, you draw up the finished shape you want in a 3D drawing package then make the "negative" of it as a mold, add the "runners", "vents", ejection pins (often not needed for low production) and split it into two pieces. You fire off the drawings to a mold machining company that uses CNC machining to produce the mold. For high volume, molds are usually made of steel.

Two-part reactive-cure polyurethane can be molded in 3D printed molds. PU doesn't stick to anything voluntarily, so you generally don't have to fuss too much with mold release compounds. If a mold release is necessary, there are spray-can types that will work. I would consult the manufacturer of the material being used for recommendations for release agents.

Epoxies can be used, but because they are normally very hard and rigid, mold design must be done very carefully or you'll never get the part out of a rigid mold. Epoxies do stick to things and mold release treatment may be crucial. Sometimes molds are made from flexible materials such as silicone, though this leads to lower precision.

Molding with any material that requires mixing of two parts is not trivial if good quality is required. The parts must be mixed thoroughly. Sometimes "static mixer nozzles" are used on double-barreled cartridges. Static mixers work well, but if you don't have multiple molds you'll use one nozzle per part and waste a nozzle full of material for each part. If the material is mixed in a pot it is usually necessary to de-gas (remove air bubbles entrained during mixing) it by placing the mixture in suitable chamber and drawing a fairly strong vacuum. If you entrap air as the material is poured into the mold, you will get imperfections. Filling the mold while it is under vacuum may be necessary and isn't especially easy.

"Potting" usually involves putting something into a container that serves as a mold and usually remains permanently. Silicones and epoxies are popular for potting. Again, degassing is often necessary for best results but somewhat less demanding because you can often de-gas the mixture, pour it into the potting "box" at atmospheric pressure then put it in a chamber to de-gas if necessary. The second degassing often isn't necessary.

When a potting compound such as an epoxy is used, you must consider the stress that can be put on things during curing. Potting compounds can expand due to heat of reaction, then contract as curing completes and this can put a lot of stress on parts. "Toughened" epoxy can be less stressful to parts that are potted, but it is generally quite expensive and there aren't many choices.

Getting anything to bond well to cable insulation to assure a water-tight seal is not at all easy. Sometimes certain solvents can be used to enhance bonding. Tetrahydrofuran (a cupcake with a cherry on top, as I describe it) can be useful for urethanes. But then you have to consider where the solvent goes. Eventually it will diffuse out, but that can take a very long time.

===
Probes like that in the photo should be excited with AC with zero DC component. If there is any DC component the metal on the probe will fairly rapidly dissolve - not to mention that DC will also result in polarization of the electrolyte which cause measurement error.
 

Ajk Tek

Joined Jul 9, 2017
11
I have had luck using parafin wax as a release agent for 2-part epoxys. You can use just about anything for the mold that is close to the size of your sensor and cover it in parafin wax so the epoxy can easily break away when hardened.
 
Top