water level sensor

[John HaineParticipant @johnhaine32865]

What is the tube diameter? I use these a lot in clocks, have a 10mm gap, you may be able to get bigger. 5V in through 270R resistor, logic level out.

[duncan webster 1Participant @duncanwebster1]

Posted by Nick Clarke 3 on 04/03/2021 09:36:45:

I seem to remember an article in ME by the later Roy Amsbury which used sensors screwed into the boiler which when covered by water indicated to high and too low – It was when I first began taking ME as a teenager so early 70s.

Of course you wouldn't have a visual indication then.

If I get chance later on I'll see if I can dig it out.

I've got that article, but I'm trying to avoid probes and continuity.

I've got one of those sensors linked by John, I'll give it a whirl later, but it's #2 son and #1 grand daughter's birthdays, so today will be taken over shortly

[Martin KyteParticipant @martinkyte99762]

Often water guages have diagonal striped cards/plates fitted behind them. The water in the tube reverses the stripes. For an electronic guage sensor if the stripe was half and half vertically it should appear to swap sides changing from say white to black which should be detectable with an optical sensor.

Alternatively use one of these.

**LINK**

regards Martin

[mechman48Participant @mechman48]

Posted by duncan webster on 03/03/2021 22:23:00:

Posted by Grizzly bear on 03/03/2021 18:22:36:

Duncan,

Would a solid float work, sitting on top of the water?

Good luck, Bear..

That would be neat, but what would I make it from bearing in mind it will be 4mm diameter or less.

The simplest method works best, have a look at…

4mm Diameter Solid Delrin Polyoxymethylene (POM) / Celcon Plastic Balls Simply Bearings Ltd..

work on the acronym 'KISS' . Sometimes projects are over engineered where 'simples is as simples does' to closely paraphrase a line from a well known movie.

George.

[Ian PParticipant @ianp]

Posted by mechman48 on 04/03/2021 12:05:20:

That would be neat, but what would I make it from bearing in mind it will be 4mm diameter or less.

The simplest method works best, have a look at…

4mm Diameter Solid Delrin Polyoxymethylene (POM) / Celcon Plastic Balls Simply Bearings Ltd..

work on the acronym 'KISS' . Sometimes projects are over engineered where 'simples is as simples does' to closely paraphrase a line from a well known movie.

George.

I was going to look up the density of Acetal but looked at the ball you linked to first. Not surprisingly one line of text below the detailed specification states 'These balls will not float'. TBH that is not strictly true as they would float on Mercury at least.

Ian P

[Michael GilliganParticipant @michaelgilligan61133]

Posted by mechman48 on 04/03/2021 12:05:20:

.

The simplest method works best

.

So why use anything other than the original design [which has no added parts] ?

MichaelG.

[Grizzly bearParticipant @grizzlybear]

Hi Duncan,

"That would be neat, but what would I make it from bearing in mind it will be 4mm diameter or less."

Make it from anything that floats.

If your liquid is water, oil based liquid will float, I was thinking polystyrene or similar.

You are a man of many talents, think outside of the box, or in your case, tube.

If I'm talking rubbish, I haven't read all your posting.

Defence rests……….

Bear..

[SillyOldDufferModerator @sillyoldduffer]

How about sonar? A small piezoelectric transducer bounces a ping off the water/steam interface.

I think better to ping and detect at the bottom because I suppose the speed of sound in steam varies with pressure, which would cause the 'level' to appear to alter depending on what the boiler was doing. I think the speed of sound in water doesn't alter with pressure but could be wrong.

By alternating pings from top and bottom. it might be possible to measure water level and pressure with the same device.

Experiment required. Can an ultrasonic echo be bounced off the surface at all? If echo location works, what's the minimum distance it will work over and how accurately can distance be measured? The speed of sound in water is about 1450m/s, so pinging 10mm of water would take about 13uS, a bit fast for an Arduino.

Dave

[John OlsenParticipant @johnolsen79199]

Is this for a plant in a small boat? I wonder if in fact it would be simpler to arrange a variable stroke pump and then adjust the setting over a number of trials until the sweet spot is found where the boiler level is maintained correctly without needing continuous attention. You might think that would not be possible with the speed of the engine possibly varying, but in my not so small steam launch, I have found that (purely by chance) one of the two pumps will maintain the level correctly in the boiler for long periods without having to either turn the pump off or turn the second pump on. When I say long periods, last Monday a group of us made a trip in Whanganui that took about five hours, and the pump settings were not changed for most of that time…I had the two on briefly at the start to get the level up to a good point, then left them alone for most of the trip.

Another approach that has been used in small but man carrying launches is to have a float in the hot well controlling the pump bypass. The requirement here is that no water should be being lost anywhere else in the system, through glands or excessive whistle blowing etc. So there is a fixed amount of water in the system, and if it is in the hotwell then it is not in the boiler and vice versa. This wouldn't work for me at present, since water is escaping somehow into the bilges slowly, so every couple of hours a bucket or so from the makeup needs to go into the hotwell. I think it is escaping from the pump glands slowly.

There is a much more sophisticated system used in full size practice where you have a closed end tube connected to a column at the desired controlled water level. The tube slopes upwards from the connection point. When the water level in the boiler is at the height of the connection point, the steam above it will condense due to heat loss from the tube. Water is not a good conductor, so the top end of the tube will cool down. When the water level is below the connection point, any water in the tube will drain out, it will be full of steam, and will be hot, since any condensate can run out and be replaced with steam. So a temperature sensor here can detect the change and be used to control the pump. This works very well in full size practice, but might be harder in a small size, especially if it is in a boat that bounces around a bit.

John

[Michael GilliganParticipant @michaelgilligan61133]

Posted by duncan webster on 02/03/2021 22:29:40:

.

I'm particularly interested in the sensor. Does it just shine a light straight through the glass tube, in which case why does having water in it make any difference. […]

.

Astonishing [but interesting] ‘scope creep’ in this thread. considering Duncan’s actual question ^^^

MichaelG

[John HaineParticipant @johnhaine32865]

A neat sensor is a transparent rod with a 45* cone on the end above the fluid with the point just below the desired level. When you look down the rod if the fluid is below the cone level there is total internal reflection so the top appears light, but once the cone is submerged most of the light escapes into the fluid and the top appears dark. Use an optical proximity sensor to detect the difference.

[Nigel Graham 2Participant @nigelgraham2]

Intriguing idea, Dave!

A flat water surface is highly reflective to sound incident perpendicularly to it; but some still gets through.

The speed of sound v, sometimes 'c' in the calculations) In a medium depends on density.

Water is practically incompressible so pressure does not affect c, but the temperature certainly does by changing density, so as the temperature in a boiler is a function of pressure, c will vary with steam pressure. However, from an operating point of view the temperature at working pressure will be what matters, so any error at lower pressure is probably not important. As long as the device indicates there is water in there…

Now, the limits on depth and range are going to depend very much on wavelength (L) hence frequency f. It needs wavelengths shorter than the distance.

You give 13µs for 10mm distance – is that one-way or two-way travel time? We are looking at MHz frequencies – I don't know what upper limit can be achieved, but how are you receiving the return? On an adjacent transducer or the same one, by reciprocity? We are looking at some very fast switching speeds as well as very high frequencies…..

'

It may be possible to ascertain pressure by sound-speed through the steam, but I don't think this would be very accurate. A strain-gauge sensor would be simpler and potentially more accurate.

'

Thinking of optical water-level sensors, suppose the normal gauge-frame were adjoined by an opaque metal column with a glass cap top and bottom, between an l.e.d. and suitable photo-voltaic cell. Pure water won't show much, but if tinted by the usual tannin-based boiler water-treatment, would that give enough light attenuation with depth to be detectable by the cell?

[Robert Atkinson 2Participant @robertatkinson2]

Michael G. is correct, the sensor uses a directional source (Spot or slit) which is off-axis with the detector. Empty tube has two identical but opposing "lenses" so has little effect even if thick walled. When filled with water it acts as a cylindrical lens and spreads the light beam. This causes some light to hit the sensor. There needs to be some depth to the hole or slot on the source and detector to make them directional. A slot (must be parallel to tube axis) gives a stronger signal but less level resolution.
One trick is to use a LED identical to the source as a photodiode detector. They are cheap, work well and are spectrally matched to the source. Jut don't use a white one. Small red laser dodes are now cheap enough to consider as a source. If you have a red laser pointer it is good for visual experimentation.
I've used this in commercial products in the past.

Robert G8RPI.

[duncan webster 1Participant @duncanwebster1]

Although not explicitly stated in the original post, this is for the gauge glass on a model boiler, so it is hot, potentially 160C+. I think this rules out Martin's link, which quotes a max temp of 80C. Otherwise that seems very similar to the one I linked in the original post. I think it also rules out plastic floats. The glass is 6mm OD, 3.8mm ID, so SOD's idea of sonar would need to be very small and tolerant of heat. It's for a test boiler, so demand can vary from nothing to max, and it might not be driving an engine, or the engine might not have a built in pump, so John's fixed pump displacement won't hack it. It won't have a hot well, so that's a non starter

I've just spent a fruitless hour trying to find the slotted opto I know I've got. I've obviously put it somewhere safe. I'll have to wait until my opto transistors arrive.

As to why not just copy the original, well I wouldn't learn much doing that, and even at my advanced years that is a big part of why I set about things.

I'm beginning to think that Michael's idea of cylindrical lens might be the answer, if the water makes it splay out like a fan it would cause a step change in intensity, same amount of light in a wider beam. I'm going to rig up an LED facing a tube and have a play, but tomorrow, tired out by birthday festivities. 

I quite like Noel's capacitance idea, but it's beyond my electronic expertise. Are we talking relaxation oscillator and phase locked loop chip? If so I've just bought some of those for another project, but I can't find them either!

Edited By duncan webster on 04/03/2021 22:02:46

Edited By duncan webster on 04/03/2021 22:03:34

[Michael GilliganParticipant @michaelgilligan61133]

Posted by duncan webster on 04/03/2021 21:58:33:

[…]

I'm beginning to think that Michael's idea of cylindrical lens might be the answer, if the water makes it splay out like a fan it would cause a step change in intensity, same amount of light in a wider beam.

 

.

Using my arbitrarily chosen numbers: The empty tube splays it more than the water-filled tube; but yes, there would be a step-change in intensity.

If you can give me actual diameter, bore, and refractive index … I can demonstrate.

MichaelG.

.

Edit: Here’s a little taster

.

Edited By Michael Gilligan on 04/03/2021 22:26:26

[Martin KyteParticipant @martinkyte99762]

Posted by Michael Gilligan on 04/03/2021 22:08:55:

Posted by duncan webster on 04/03/2021 21:58:33:

[…]

I'm beginning to think that Michael's idea of cylindrical lens might be the answer, if the water makes it splay out like a fan it would cause a step change in intensity, same amount of light in a wider beam.

.

Using my arbitrarily chosen numbers: The empty tube splays it more than the water-filled tube; but yes, there would be a step-change in intensity.

If you can give me actual diameter, bore, and refractive index … I can demonstrate.

MichaelG.

That's why I suggested the virtical half and half stripe. With the water the background changes side so should be able to be picked up with a reflective type sensor arrangement mounted off centre to the tube. I don't see why a transmissive sytem should not work though.

I would think that the outside of the tube is going to be somewhat less than 160 degrees but I agree it would challenge the Farnell sensor at 80 degree max.

regards Martin

[duncan webster 1Participant @duncanwebster1]

Posted by Michael Gilligan on 04/03/2021 22:08:55:

Posted by duncan webster on 04/03/2021 21:58:33:

[…]

I'm beginning to think that Michael's idea of cylindrical lens might be the answer, if the water makes it splay out like a fan it would cause a step change in intensity, same amount of light in a wider beam.

.

Using my arbitrarily chosen numbers: The empty tube splays it more than the water-filled tube; but yes, there would be a step-change in intensity.

If you can give me actual diameter, bore, and refractive index … I can demonstrate.

MichaelG.

.

Edit: Here’s a little taster

.

Edited By Michael Gilligan on 04/03/2021 22:26:26

OD 6mm, ID 3.8mm, refractive index no idea!

[Michael GilliganParticipant @michaelgilligan61133]

Posted by duncan webster on 04/03/2021 22:41:59:

.

OD 6mm, ID 3.8mm, refractive index no idea!

.

Not a bad guess then … I chose OD 6mm, ID 4mm

Do you have a name for the glass ?

… if not, I will assume it to be borosilicate [i.e. Pyrex or similar]

I will adjust things tomorrow

MichaelG.

Edited By Michael Gilligan on 04/03/2021 23:00:57

[duncan webster 1Participant @duncanwebster1]

It's what is sold for gauge glasses, I'd guess borosilicate but no evidence

[Michael GilliganParticipant @michaelgilligan61133]

Here we go, Duncan …

Tube OD 6mm, ID 3.8mm, Material = botosilicate glass

Red light from a point-source 3mm from the tube surface

Tube filled with Air / Water / Glass

.

.

.

.

Q.E.D.

.

MichaelG.

[Chuck TaperParticipant @chucktaper]

Posted by Michael Gilligan on 05/03/2021 08:42:04:

Here we go, Duncan …

Tube OD 6mm, ID 3.8mm, Material = botosilicate glass

Red light from a point-source 3mm from the tube surface

Tube filled with Air / Water / Glass

[Snip, Snip, Snip]

.

Q.E.D.

.

MichaelG.

I'm intrigued – how, what etc??

What software is that?

Regards.

Frank C.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Chuck Taper on 05/03/2021 08:54:41:

.

I'm intrigued – how, what etc??

What software is that?

Regards.

Frank C.

.

RayLab, by Kamyar Ghandi … Running on iPad Pro 9.7”

**LINK**

https://apps.apple.com/gb/app/raylab/id710190065

… I gave it a rave review

MichaelG.

[Andy StopfordParticipant @andystopford50521]

Posted by duncan webster on 04/03/2021 21:58:33:

I quite like Noel's capacitance idea, but it's beyond my electronic expertise. Are we talking relaxation oscillator and phase locked loop chip? If so I've just bought some of those for another project, but I can't find them either!

Here's a fairly simple device for capacitive sensing, using a 555 timer and an ATMega 328 microcontroller (as in an Arduino). The schematic, MaquinaSafetyLevel.sch, is in the PCB folder, and is a KiCAD file.

https://github.com/andystopford/SCRAM

It works in the opposite sense to what you want, i.e.it switches on if the water is above a pre-determined level (to power a heater element).

The level probe is a commercially available one intended for espresso machines, though it seems to be just a stainless steel rod, with all but the last 10mm or so sheathed with acetal or something similar.

In main.ino, at line 10, I've commented it as voltage, however I think the values are arbitrary – it certainly doesn't carry 600 volts (I hope). These values were determined experimentally, and take a while to settle down; I'm not sure why, but I had to change them several times in the first few weeks of operation. They always drifted towards the 'safe' condition though, with shutdown occurring even when there was water in the boiler.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Michael Gilligan on 05/03/2021 08:42:04:

Here we go, Duncan …

Tube OD 6mm, ID 3.8mm, Material = botosilicate glass

.

.

Oops … Arthritic finger problem

botosilicate should read borosilicate

MichaelG.

[duncan webster 1Participant @duncanwebster1]

Michael's ray plots are extremely interesting. While I get my head round them, I've done what I should have done right at the beginning, got off my backside and done some investigating

first photo shows a 5mm LED, no glass, just to get a baseline

then with just the empty glass tube, note how the LED appears much narrower, and there appear to be 2 red edges

 

Then with water in the tube, the image fills the tube and seems much brighter, at least to my eye

 

 

I'll have to wait for the phototransistors to arrive now, but at last I'm beginning to see what is going on. I have a feeling that if Michael widened the beam slightly on his with-water plot he'd get total internal reflection on the inner glass/air interface, which might explain the side bands on the second photo.

Now to delve into Andy's capacitor scheme.

Thanks to everyone who has contributed, quite a bit learned. And the good news is I've found the PLL chips, but not the slotted opto

Edited By duncan webster on 05/03/2021 11:31:46

[Author]

Posts