I have a project that needs waterproofing. The easiest way to do that is using some form of Potting. The P0 and P1 has a metal casing around the assembly, but looking at the internals of a P0 it looks like potting should go fine?
This looks like a tempting solution: https://vimeo.com/118732965
I’ve only seen @bko’s comment in this thread but would love to hear from others that have tried potting their projects.
That Technomelt and Moldman look great but they don’t talk about the cost of having molds made–I have spent a lot money in the past on injection molds for parts so I would find someone who has used it and ask about total costs, including all the costs for molds and revision to the molds.
You can pot parts yourself with RTV silicone for prototypes or low-volume production. It is a bit messy and you generally need a vacuum chamber to get all the bubbles out, but it can be done. As I said in the other thread, operating temperature is key and working through the heat transfer numbers in advance would be a good first step.
Once again, if the part does not need to meet IP65 submersion, I would recommend you try a conformal coating first and see if that meets your needs since it is so much easier.
If you look at most consumer items that are meant to go out in the weather, they are generally made with clam-shell cases with silicone gaskets around the edge. If your product is automotive, they like to use potting, but understand the design life of automotive electronics is only 10 years and the repair model is replacing entire parts so potting is OK.
Thanks @bko!
My hope was to solve both the enclosure and waterproofing in one go. So the P1 can certainly run warm at times, but would it go above 90C? Potting will prevent any heat to escape through air, so some sort of heatpipe would certainly be required…
Maybe I can get away with potting just the part around the connectors? I think I’ll go for that approach first. I’ll do some experiments and see how that goes.
Have a look at this from MG Chem. which Mouser carries:
The key to managing the thermal problems is area and volume–the potted assembly has to be significantly larger (i.e. have more surface area) to radiate the heat from the inside to the outside due to the thermal resistance of the potting material.
I think experimenting is a good first step–good luck!
Those materials looked great, but unfortunately they’re “Shipping Restricted” :-/
I guess I’ll have to find an European distributor for something similar. I’ll post results back in this thread when I get the experiments going.
@bko do you have a suggestion as to how to increase the temp as much as possible in the chip? Calculating PI with X digits and constantly sending the result to a HTTPS server maybe?
@jenschr, if you can drive 5 external LEDs (via GPIO) with resistors limiting the current to 20ma each, then you will be near the 120ma max package current for the GPIO. This should produce a few hundred mW of power to heat things up.
Killer video! The examples they provided opens up some new possibilities for me. I love how they do the clear and then black plastic to provide the finished look and to add wire stress relief.
You can possibly consider making molds using a 3D Printer like the Forum1 where they are showing people how you can create a 3D printed mold for injecting liquid plastic into them. This should work fine for the Technomelt material also.
@peekay123 I’m now driving 7 LEDs and I’m not seeing much of a temperature increase? Here’s a picture of my setup. I have a custom PCB based on P1 that has the RGB-LED and 5 white LED’s with low value resistors for approx 20mA. Since I’m not seeing much of a temperature increase, I’ve added a couple external blue LEDs driven from some of the broken out IO pins and I’m still not seeing a significant temperature increase. Can it be that the P1 module can pull more than 120mA?
On top of the module is about 8mm of clear epoxy ( Electrolube Water White, UR5634RP250G). No epoksy covering the bottom of the PCB yet since I wanted to see if any heat escapes there (from the ground pins on the P1 module). It doesn’t look like that. I’m quite happy with this for now as I doubt my device will ever pull more power than this from any pins, so I’ll do the second layer of black potting tomorrow that will also cover the bottom of the PCB.
Note that potting is going to reduce wifi performance; the increased permittivity of the potting compound will detune the antenna (the resonant frequency will reduce, likely significantly, moving it out of the wifi band).
This may be important, depending on your range requirements. Conformal coating, being thinner, will affect this less - but it will still affect it.
@hfinnes Are you sure? Won’t that depend on the material? I’m not seeing any signs of poor reception and as you can see - the antenna of the P1 is out towards the edge of the board so it’s not much material blocking it.
I’ll make a couple more boards this week and compare wifi performance. If it’s poor I’ll just break out a SMA connector instead, but I’m kind of hoping this will be sufficient as it saves parts & assembly.
It does depend on the material, and the higher density the material the bigger the effect. There’s a fairly large jump from air to anything solid though.
The antenna can be tuned to work in a higher density medium, but that means changing the pattern itself and possibly the match components within the metal can of the module which generally should not be attempted without a network analyzer (plus you invalidate any wireless certifications you may have had previously).
Without test gear the only thing you can do is check RSSI at extremes (ch1/ch11). The problem is with wireless is getting a repeatable test setup is near impossible as the smallest physical change (or proximity of a person even) will alter results.
Any specific cases where somebody else you know did this exact same thing and had wifi signal issues?
Yep. Commercial product (imp-powered outdoor sensor node a few years back).
There are potting compounds with lower dielectric constants (which is directly related to permittivity), eg http://www.epicresins.com/ElectronicsPotting/rfdevices
See other advice about potting and RF all over, eg:
“Radio frequency” section of http://www.pottingsolutions.com/my%20site/Technology/potting_hints.htm
https://www.eevblog.com/forum/projects/antenna-in-a-non-standard-enviroment/
https://www.lsr.com/white-papers/antenna-matching-within-an-enclosure has some good general tips, but it’s more about production where you can and do tune your antenna appropriately.
To follow up on this one. We ended up taking a chance and it worked. We produced injection moulded boxes in ABS. We then put the electronics in and filled the box with Electrolube’s “UR 5048K5K”.
The top of the box is split in two due to our waterproof connector’s irregular shape. We also have 3 buttons behind the connectors the required special care so they didn’t fill with potting. The process is a little cumbersome to perform, but the result became really nice. The device isn’t 100% waterproof, but it can withstand splashes easily.
The RF properties were not affected by the potting in any measurable way and we also have an external antenna connector that would get around that.
