Building a Cheap, Portable Water Level Recorder for Seiche Monitoring

[DaveC426913]

TL;DR Summary

Looking to make a DIY sensor that records water levels

I want to build a cheap, portable water level recorder so I can record the "seiche" on my lake (and other lakes).

I don't want to use fancy components like Arduino boards. And I want it to be portable (and cheap enough that losing it to water, weather or thieves won't be a heartbreak.)

There's two parts: the float plus stilling well, and the recorder. I can manage the float and stilling well. I'm hoping to find some simple, elegant way to rig up a recorder.

The most obvious way is to have some motorized drum with paper, but I'm looking for more off-the shelf parts.
It doesn't really have to be sophisticated - for example, there's no reason why it needs to rotate 360 degrees in 24 hours. I don't plan to have it record that long, and I can always calibrate the graph to the time however is necessary.

I'd prefer it to be self-contained electrically (i.e. not needing a power outlet). (If I have to have power, I might just use a ten dollar wall timer for lamps.)

Really, anything that moves or rotates slowly and consistently ought to be usable to rig up with a marker attached to the float.

I see the whole contraption as being no bigger than a foot around and 2 feet tall (to reach into the water). (The min/max is less than six inches).Can you suggest some types of devices - ideally, battery, or sunlight, or spring-powered that will be handy for building a moving graph?

 

[DaveC426913]

Thinking laterally, just for fun, a camera set on time lapse will do a fine job of recording directly, without need for motor, drum, paper or stylus.

 

[russ_watters]

Does it need to record more than one cycle, and does it need the variation over time or just the extrema?

 

[berkeman]

Summary:: Looking to make a DIY sensor that records water levels

for example, there's no reason why it needs to rotate 360 degrees in 24 hours.

To record one 24-hour cycle, you could probably base it on a wind-up clock mechanism. Fix the hour hand in place, and let the body of the clock rotate. Put your paper disk on the clock face, and attach your pen/float part to the fixed hour hand part.

 

[DaveC426913]

Does it need to record more than one cycle, and does it need the variation over time or just the extrema?

(I think I know where you're going. Min-max thermometers have a simple mechanism with two plugs that get lodged at the min and max points. An analogous contraption could be set up for water levels.)

My primary interest is in the timing of peaks (because the period is directly related to the shape of the body of water). But there may be several seiches overlapping, so seeing the variation over time would highlight them all.

I would have to record more than one cycle, else I'd never know if it's a pattern. But the cycle for huge lakes like Ontario is only about an hour. So I don't really need to record as much as 24 hours.

 

[DaveC426913]

To record one 24-hour cycle, you could probably base it on a wind-up clock mechanism. Fix the hour hand in place, and let the body of the clock rotate. Put your paper disk on the clock face, and attach your pen/float part to the fixed hour hand part.

I'd thought of using a clock, but I'd imagined a drum gizmo sitting on top of it with a cylinder of paper. It had never occurred to me to just use a horizontal disc of paper on top of the clock face. That would simplify things dramatically.

That's elegant.

 

[DaveE]

I might try building it up from eBay kits. An ultrasonic distance measuring kit, maybe a digital volt display, or a cheap data logger. That stuff can be incredibly cheap.

Anyway, the idea is you make a leaky tube with an ultrasonic range meter mounted at the top. Arduino, etc. doesn't have to be complicated or expensive if someone else built and programmed it for you. Mostly I suggest this because I think it would be fun.

OTOH, if you can take pictures, then how about a leaky tube with a float, a string up to a pulley at the top and a gauge painted on the outside of the tube. Like you see on farm water tanks? Can't get cheaper than that.

I think the real issue here is how do you want to log the data more than how to measure it.

 

[DaveC426913]

I think the real issue here is how do you want to log the data more than how to measure it.

I expect I'll do that manually, taking readings off the raw chart.

This is just a fun experiment - a follow-up from my serendipitous discovery last summer. My measurements and recording techniques were clumsy in the extreme, and I'd like to fix them.I'm reading several of Tristan Gooley's books on Reading Nature's Signs and am getting fascinated with exploring such things the more I sail and the more I hike.

I might try building it up from eBay kits. An ultrasonic distance measuring kit, maybe a digital volt display, or a cheap data logger. That stuff can be incredibly cheap.

Hm. Now you've got me thinking...

The device is for my own use, but if I built it more robust and with continuous reading, I might be able to set it up at my yacht club permanently and score some brownie points.

 

[anorlunda]

You may be interested to hear that there can be seiche waves in the thermocline in lakes and seiche waves in salinity inn oceans.

Measuring those could also be DIY fun.

 

[jrmichler]

By the time you put together a mechanical chart recorder, and find a way to power it, you might be better off to use an Arduino setup. EnviroDIY (https://www.envirodiy.org/mayfly/) makes a board specifically designed to applications like yours. It has built-in provisions for a real time clock, microSD card, 10 bit and 16 bit A/D, solar power, and Li-ion battery. They sell their boards and starter kits through Amazon: https://www.amazon.com/dp/B01FCVALDW/?tag=pfamazon01-20.

Some gas gauge sending units are rated for use in water. A search for water level sending unit found this, plus others: https://www.amazon.com/dp/B079L7YL2W/?tag=pfamazon01-20.

One caveat with the Mayfly - both the 10 bit and 16 bit A/D inputs require sensors with output impedance less than about 30K ohms for best performance. At 100K ohms, input noise is significant. Above 300K ohms, the input noise and zero drift make it nearly unusable. Data is recorded to the SD card in text format, and you have to remove the card from the Mayfly to read it.

Adafruit (https://www.adafruit.com/category/63) sells displays that can be interfaced to Arduino to make a real time strip chart display. I have no experience with these displays, so cannot make a knowledgeable recommendation.

 

[DaveC426913]

By the time you put together a mechanical chart recorder, and find a way to power it, you might be better off to use an Arduino setup.

I might. I just contacted my buddy who also sails at the club, and he plays with Arduinos and Raspberry Pis and ESP8266s.

Although, I've spent my career programming. I'd like a break to work with mechanical stuff.

 

[Tom.G]

For recording you could use a barograph.
https://www.ebay.com/b/Barograph-In-Collectible-Barometers/14021/bn_7022178117

IIRC, there are also recording thermometers.

 

[Baluncore]

Maybe a type of kymograph, made by sticking small magnets onto a vertical axis steel drum. https://en.wikipedia.org/wiki/Kymograph
The magnets are guided into a position determined by the water level. At the back of the drum they are gathered by a moving 'V'-collector, to be repositioned.

Alternatively I think you might reverse engineer a 'Magna Doodle' like technology. https://en.wikipedia.org/wiki/Magna_Doodle
As the disk or drum rotates to the left, it is written by a magnetic stylus on the right in an area that has been erased earlier. The display would always be for the most recent period, which would depend on the speed of drum rotation.

 

[anorlunda]

By the way, it will work best if you create two instances of the recording device. Install one on each side of the lake.

Below is an extreme example from Lake Champlain (my favorite lake) during the passing of hurricane Irene in 2011. The three plots show lake levels north, central, and south on this 111 mile long lake. The level did not return to the initial value, because there was a lot of rain in addition to the wind driven seiche. Sorry for the fuzzy pictures, the vertical scale is 93 feet to 98 feet.

I remember that example well because I rode it out on the lake at anchor without problem, and this graphic was from my blog. A more extreme example was the day when Lake Erie was 27 feet higher in Toledo,OH than at Buffalo, NY.