Public Lab Research note

  • 3

Riffle Water Bottle Enclosure: rubber stopper w/hole + silicone sealant: > 8 weeks!

by donblair |

donblair was awarded the Empiricism Barnstar by warren for their work in this research note.

[photo credit: Liz Barry. Shocked cereal box credit: D'Ignazio family]

We've been experimenting with various enclosure designs for the Riffle water monitor prototype -- our driving question being:

What's the cheapest, most accessible, effective way to keep monitoring electronics dry underwater?

One answer is: especially if you only want to measure temperature and conductivity (and, perhaps, turbidity), and log it locally inside the enclosure to a microSD card, then a 500 mL water bottle seems to be a fine choice, and pretty much ubiquitously available, world-wide:


We've recently been playing with simply using screws through the cap, rubber washers, and nuts to seal up the conductivity probes, and it seems to work nicely, for at least a week or two. My own thinking is that for a simple conductivity & temperature datalogger, this approach will be a nice, effective one for most folks to use, with the parts readily available almost everywhere.

But we've also been interested in trying other enclosure designs, for sensor probes that require e.g. cables to pass through the enclosure (some turbidity sensor designs might require this, for example).

On August 30th, I tried a design that had come up in conversation with @mathew: using a size 3 rubber stopper with a hole in it ...


... with the rubber fitted snugly inside the bottle mouth. I placed some wires through the rubber stopper hole, and sealed the hole with silicone sealant:


I placed this in a bottle, and then about four feet underwater in the Charles river on August 31st -- and only retrieved it this last Friday -- over eight weeks later. @liz was present to witness the slimy, smelly bottle right after it emerged from the Charles:


It was still airtight -- it hadn't leaked!


Will aim to deploy some bottles this week with the screws-in-cap option ...

sensors water-quality riffle open-water bottle enclosure



Nice results, real example – 8 weeks is non trivial.

Love it! Maybe we should offer these parts in the store for early experimenters? Just the stopper and some sealant? 2.8oz tubes are for <$4 on amazon:

@tonyc, what do you think? If we put out on the lists as a special maybe we could gauge interest?

I'm not seeing any bulk sources that get us a price break over the price on amazon, so I don't really see a reason we should carry it over just linking people to the product there.

having kits carry an item makes sense if we get a huge price break, or if it is an item that is hard to find locally. These stoppers may be something worth bulk purchasing, but the sealant looks tougher.

I also think we could bulk purchase silica desiccant and sell in smaller qty, to throw in with the Riffle.

also, right now this is, like my earlier stopper example, a sealant strategy without sensors to go with it.

@Donblair has stuck to his guns on screws-through-the-cap conductivity sensor designs as the most accessible to assemble. We sourced some 0-80 machine screws and o-rings that should work better than previous through-the-cap screw designs. I believe @Donblair put some out in the field or is working on those.

I'll be bringing silicone and stoppers to the barnraising, we can play with both strategies and some sensor prototype ideas for stoppers. My impression is that we want a 1.0 riffle kit to focus on screws-through-the-cap system.

Yes -- at this point it might very well be possible to do temp, conductivity, and turbidity while only needing to put screws through the cap, and this approach does seem very elegant to me!

It's good to know that we can put wires through the sealant and still remain waterproof for long periods of time -- this enables a simple design for a depth sensor that would involve a long wire sticking out of the cap ... which is probably not for v. 1.0 of the riffle, but could be very useful down the line ...

@mathew -- we used the 0-80 machine screws on ~ 20 sensors in Bogota, and had two issues: they were on the tiny side re: manipulating them, and the o-rings seemed to pop out from under the screw head when we applied significant pressure in screwing them in. So we might want to play with a few options ... but still serviceable!

I was worried about those o-rings... the silicone, while durable, is flexible and doesn't stick down well.

Next option is to try poking through a sheet of rubber on both sides as you suggested. I'm thinking a bike inner tube (also because I want to try putting two screws and a presta valve through the cap)

Would they be easier if the screws were longer? They are so cheap we can easily include a screwdriver in the kit.

You must be logged in to comment.