Hello everyone!

I am planning on starting a bigger electronics project which involves creating my own weather station from scratch. I am still in the planning stages of it, but I have compiled a list of what I want the station to be able to measure and I have also decided on the shape & concept it should have. I realized that I would like it to be outdoors, so it would have to be waterproof. I want my station to be completely modular, with each sensor unit physically looking and behaving like a lego block (so if I want the station to measure something new, I'd just add a new block-shaped module on top of the others). For this connection to work, I'm planning on using a 4-pin connector (VCC, GND, SDA, SCL), alternatively 5-pin (VCC, GND, SDA, SCL, Analog stuff) which would connect between the block modules.

So, to my question: I need to find a suitable connector. I would ultimately like the block-shaped modules to connect to each other (again) like legos do, aka. there would be a male connector on top of each module and a female on bottom of each one. The problem is, I have spent hours looking online without finding a suitable connector. If it's going to be on top of each module and be waterproof, it can't have screws or clips because the user can't get underneath to unscrew or unclip. The user should just be able to pull the whole module out. So basically tl:dr; is there any 4 or 5-pin connector you can purchase as a hobbyist/student that is waterproof without requiring the user to unscrew/unclip/unwhatever it?

I haven't made any drawings of the weather station idea yet, but if it's difficult to understand my ideas, reply and I'll make a drawing for you. The backup plan is making the connections go on the side of each module instead of on the top and bottom so the user can reach the connectors.

 

Welcome to AAC.

[TLDR: This turned out to be a lot longer than intended, and it doesn’t directly answer your question, but I do think is it important and it comes from many years of working on the design of various systems. On the other hand, please feel free to ignore it if you don’t find it helpful. I, too, could not find a single example of a waterproof board-to-board connector and I don‘t think one exists since it doesn’t seem to solve any engineering requirement you are likely to find in the real world.]

A drawing would help, to be sure. But, as far as I understand your requirements, you want a stacking, non-locking, board to board, waterproof 4 or 5 pin connector. I am afraid you are not likely to find such a beast for a variety of reasons.

Let’s back up and see what your requirements are, not your attempt at meeting them. I think you are confused your first about how to solve a problem with a requirement. After all your searching you have discovered that your imagined solution doesn’t exist, and there must be a reason.

First, you say you want to have this be waterproof, yet you want a board-to-board connection. How will the boards be waterproof? Normally, you don’t expect boards to be directly subjected to water. Instead you put them in a waterproof enclosure and if reliability is a top priority, use a conformal coating on the boards to limited potential moisture Ingress. I can’t recall ever seeing a PCB with conformal coating, directly interacting with water though I imagine there must be some instance of it because of an absolute size requirement or something else.

Nonetheless, everything comes with a cost, both in terms of dollars and time. Con formally coated boards are very hard to maintain, for example. And ensuring that a connector body, which necessarily represents a penetration in the coating, doesn’t allow ingress is not trivial. Particularly over time, and if the board remains submerged in or covered by water.

So what are your actual requirements? It seems to me, in the absence of other information, it goes something like this:

1. Develop a modular design so sensors and other functional elements can be added without special accommodation for the type of module. This includes auto-configuration capabilities such that there is ideally no user interaction required outside adding a module.

2. Establish a common bus for power and communication signals that allows for arbitrary expansion. This includes the ability to expand by “stacking” modules of any type.

Those are the base requirements, in pursuit of satisfying them there second order are software and hardware requirements, and preferences which shouldn’t become requirements unless they really are. So your desire to use a “lego” model may or may not be a requirement. The engineering tradeoffs are part of working out a design that satisfies the primary requirements while being both feasible based on cost and actually effective.

It is a nice idea that they should stack ”like lego”, but is this really a requirement? This is a question since you are the designer, it could be. If so the cost of engineering this idea becomes part of the cost of accomplishing the project goals. Is there some part of the lego-like behavior that is more important than others? (e.g.: mechanically fitting together with some other electrical path, etc.).

Let’s first assume the lego bit is a desirable outcome but not a requirement in the sense of the two items above. We can start engineering a solution for the primary goals, at least as part of the first prototype, that lets you test your theory that stacking is actually desirable. Very often executing an idea that seems good in the abstract makes problems in actual 3D space much more evident.

If we assume the lego concept is very important to you, why not start with ordinary stacking .1” pitch headers. These are common, easy to apply, and will let you test nearly every other aspect save the waterproof part. At the same time. You can conspider the mechanical aspect of waterproofing. The stations I am familiar with depending on the housing to protect the PCBs, not inherent waterproofness of the PCBs themselves which is a much harder problem.

Don’t forget that you should be making prototypes that are a sort of path of least resistance in terms of testing as many parts of the design as possible while not trying to accomplish everything at once. If I were doing this, I think I would start (if not keep) a standard NEMA enclosure, already engineered to be waterproof eliminating, at least initially, the need to take on this very non-trivial engineering challenge.

I might also begin with a card cage physical architecture. This allows you to test the concepts of modularity and zero-config. Also, don’t forget that the PCB itself is unlikely to be the ideal place for an associated sensor, so something on the bus needs to allow for remote sensor connection, possibly with a connector per slot. Since not all modules will require remote sensors, you might have classes of slots to cut down on wiring and connectors.

So if something is sensorless, say, a GPS module used for location and time that doesn’t need a remote antenna, it can be placed in a slot that doesn’t have access to a connector for a sensor, while something like a wind speed module, that needs to be connected to an anemometer, which require a slot with a connector broken out from the pins on the bus connector for that slot. (As an aside, having the slots with addresses the boards can read would be a nicety for any required user interaction since the module could report its own location. Possibly, a Dallas Semiconductor one-wire ROM chip would be a direction).

Even if ultimately your lego idea can be done, given the difficulty of even finding a connector, and the question of other waterproofing, I think your first prototype should kick it down the road. There are many other problems to be solved, and as different requirements appear from practical engineering, you won’t be working around the limitations it presents without having been able to make a cost-benefit analysis.

If I were doing this I would be very concerned about specifying the bus and making sure the signals needed were there rather than the mechanical realization of the idea. A backplane arrangement, for example, allows for expansion to cover unforeseen needs by adding pins that are “reserved” while the current state of your stacking arrangement doesn’t make that at all easy.

Sorry for the length of this, your question provoked a bunch of stuff related to design that I spent a lot of my career woking on with others. I could say a lot more about the process, but I will just say that design is iterative. Getting stuck on satisfying something that hasn’t been vetted as part of the larger system can lead to a dodgy outcome that doesn’t do anything well, not even the thing you worked so hard to include.

 

There are very few connectors where the chassis or pcb mount male will mate with the chassis or pcb mount female. I can think of DIN41612 and D-type. The D-type seem the most likely to have some degree of waterproofing.

 

One option would be to have a female pin header on each side of the circuit board, a "goes-in" on one side, a "goes-out" on the other. These would be positioned so that an "expansion port" on each side of the housing could be covered with a rubber plug or even a layer of tape.



When a new module is to be added, the rubber plug or tape would be removed, a back-to-back male header plugged into the female header on one side, and plugged into the other block when it's brought into position.



Samtec makes headers of any length to order at a surprisingly affordable price.

To improve this idea and make it a little less finicky, female headers are avaliable where the male header pin passes through instead of bottoming out. This way, the exact distance between the connectors (and boards) isn't so critical; get the back-to-back male headers a little longer than needed, and the excess length will just pass through the female header.

Doing it with a removable link like this means you don't have pins sticking out that you somehow need to waterproof and protect.

 

@Ya’akov The length of that message was nothing that bothered me, I really appreciated every single word of your feedback, especially because this is a hobbyist project and you obviously have more experience than I do. Also, thank you for the warm welcome! Sorry about the confusion though! I want the PCBs to be contained in "lego block"-like modules, so the modules would not just be the PCB. That way I am hoping to make it waterproof without requiring special PCBs. Here is a drawing:

(I realized the scan is kind of blurry - blame my notebook. I hope you can read it anyways)
@Jon Chandler Yeah, the connectors I want are going to be male-female of some sort and your idea of just using the most basic ones are something I had thrown away - but you gave me a new perspective. Thanks!

@Ian0 Thanks for your input! I sadly could not find a DIN connector that is waterproof and has as few pins as my requirement (4 or 5). I know I could always leave out a few pins, but I want as clean as design as possible. A D-connector might work though! But, I am not a fan of its footprint - it will fit but is a rather big connector.

 

Open the hood of your car, get part numbers off the Deutsch connectors that look like close matches. That will be a good starting point for your research. These connectors are highly configurable, functionally infinite number of configurations. You usually see them as board-to-wire plugs but the plugs can be ordered as male housing with male pins or male housing with female pins; female housing with male pins, or female housing with female pins, so I don't see why they couldn't be used as board-to-board connectors. Same goes for the Military style round connectors from Cannon and Amphenol. There is also the Harting connectors (Google harting HAN) and their more affordable equivalent here.

But I would scrap the idea of waterproof connectors and instead waterproof the entire module. You can have stackable modules with rubber gaskets (o-rings are best) between them and a lid goes on top of whatever module is on top. If you want to add a module, unscrew the lid, put the new module where the lid used to be, and screw the lid on top of the new module. I don't see why this is a problem. You mentioned the user not being able to access the bottom of it? Does not compute. Go look at a DB9 null modem, they are stackable, same concept.

 

@Ya’akov The length of that message was nothing that bothered me, I really appreciated every single word of your feedback, especially because this is a hobbyist project and you obviously have more experience than I do. Also, thank you for the warm welcome! Sorry about the confusion though! I want the PCBs to be contained in "lego block"-like modules, so the modules would not just be the PCB. That way I am hoping to make it waterproof without requiring special PCBs. Here is a drawing:

(I realized the scan is kind of blurry - blame my notebook. I hope you can read it anyways)
@Jon Chandler Yeah, the connectors I want are going to be male-female of some sort and your idea of just using the most basic ones are something I had thrown away - but you gave me a new perspective. Thanks!

@Ian0 Thanks for your input! I sadly could not find a DIN connector that is waterproof and has as few pins as my requirement (4 or 5). I know I could always leave out a few pins, but I want as clean as design as possible. A D-connector might work though! But, I am not a fan of its footprint - it will fit but is a rather big connector.

Based on the drawings my first inclination would be to make the connection from one module to the next gasketted. I would probably consider a card edge connector.on the top of earch module and a mating PCB on the bottom to mate with it.

View attachment 270288​

The waterproofing would rely on the gasket, not the connector.