EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Sredni on April 12, 2024, 01:33:37 pm
-
I want to make several *cheapo* moisture sensors, using stainless steel bolts (AISI 136) connected to the solid copper conductor of a Cat5e cable.
(Let's say this is a given and no other better solution is possible).
What is the best way to connect the cable to the steel bolts?
I can turn it around the bolt shaft and use a nut to keep it in place, but I need a connection that won't allow for the oxidation of copper.
I can try to solder the surface copper, use enamel, hot glue, silicon, on the connection.
The exposed part is the thread of the bolt and it is cemented in a concrete slab.
The entire sensor will be underground, and inaccessible once installed. There won't be any mechanical strain or vibration . The sensors just need to sleep there, undisturbed.
What would you do?
-
id use a solder lug with a hole suit the bolt held in place witth a nut
-
Sufficiently high contact pressures should result in a gas-tight interface between the two materials that should prevent oxidation at the interface. This is the foundation of all crimp connections. A properly tightened lug connection works similarly. Dielectric grease and covering with glue-lined heatshrink are options for further protection against corrosion.
The bigger challenge here is probably going to be the thin solid copper wire, which will be very susceptible to mechanical damage during assembly/handling that can precipitate a later failure. Strain relief on the wire-to-probe connection is key. If you really want to use solid cat5, I would suggest splicing it to a a pair of larger gauge stranded wires that run out to the sensors. You'd want to keep that splice point above ground, and/or seal it very well to keep moisture from wicking up inside the jacket of the cable (again, glue-lined heatshrink is great here). Or you could use something like speaker wire, which is typically 20AWG or larger two-core stuff that is also very cheap and skip that step.
-
A solder lug would be my first choice, but the copper could also be silver soldered directly to the stainless steel bolt.
-
Thank you all. So far, very helpful advice.
Yes, I had thought about strain relief by coiling the wire, but the idea of adding a piece of multifilar wire and then soldering that to the lug is way smarter.
I think I'll go with:
Bolt->Lug->soldered flexible copper wire>cat5 solid copper (coiled to give strain relief)
Then I can insulate everything with enamel, hot glue, and heat shrink.
-
when I made my conductivity tester for water i threaded the stainless rods and put a crimped lug on it with a washer and nut and used a thin wire. for the junction bond I used silicone grease, but for the seal....
And the silicone is maybe not good, and probably not hot glue, it lets water vapor through. you want some moisture tight RTV, typically oven cured, for optimal performance.
-
since this is a rod, can you put 2x plastic electrical boxes with water tight rubber glands (the cable gland you tighten) so then you can fix the rod to the small box with 1 gland, and have strain relief on the wire with the other gland
and cat cable is not good because its loose in the sheath, you want like a rubber cable thats molded on the wires (not a loose tube)
this way you have full service access, and no glue of any kind, since its held together with o-rings
you can like add something inside to clamp on the rod (plastic clamp) to make it more durable
-
and cat cable is not good because its loose in the sheath, you want like a rubber cable thats molded on the wires (not a loose tube)
Unfortunately the choice of the cable is a constraint linked to the "cheapo" spec 😁
A better choice would be two separate conductors to be kept apart to reduce cable capacitance or maybe even three conductors and use one as a guard. Even better, make the sensor digital and use whatever cable. That would be smart.
But I never said to be smart. ☺️
In fact I need to make AC measurements and cable capacitance significantly alters the reading. I mean even 30cm (a foot) of twisted pair can change completely the measurement of high resistance at the other end. On paper and in simulation , with a bit of math, I can get reasonably accurate measurents up to 8 meters (10 yards). I need to go only as far as 5 meters.