Help understanding how my "magic" siphon works....

[fxfocus]

I have a fish tank into which is flowing a steady stream of water at around 2L/H. I set up a submerged pump controlled by a float switch to remove water at the same rate as it was coming in and so maintain a constant water level. The pumped water is discarded to a drain via a 6mm flexible tube. Problem I ran into was that when the pump turned off the water continued to drain from the tank by means of the siphon effect. I had the idea that making a small hole in the tube just above the water line would allow air to be sucked in and break the siphon. Providing the hole was small enough any water escaping through it when the pump was running should be small potatoes and so could be ignored.

Sure enough when the pump turned off air was sucked in but I was surprised to see the siphon continue to run even with a stream of bubbles mixed in. But then came the big surprise, while the siphon was running the water level remained constant to within a millimetre..! Since the water level wasn't rising the float switch and pump were never activated.

See the siphon in operation

HOW DOES IT WORK?
My experiments indicate the distance between the hole and the waterline is critical. A very small change in water level causes the flow to increase or decrease to compensate. Moving the hole up or down (simulating a change in water level) by only a millimetre or two results in a significant increased or decreased flow. I guess it's the venturi effect that causes the air to be sucked in but how is it controlling the water level? My theory is that the length of pipe between the water level and the apex of the siphon is made longer or shorter as the water level rises or falls causing the pressure differential to vary. Be interested to here what others think...

IN SUMMARY
There is no new science here, of that I'm sure. In all my years of fishkeeping I thought I had seen every type of siphon under the sun but I've never seen anything like this before. I'm sure it's practical as it's been running on one of my tanks with no issues for nearly two months. Is anyone out there already familiar with this application of a siphon?

 

[Lnewqban]

In the video, where is the tiny hole located at?

 

[Arjan82]

Hi, sound like a fun application! Can you perhaps make a sketch of the setup? I'm not entirely sure if I understand all the details.

 

[anorlunda]

Normally, the air hole in a siphon break is located at a point higher than the water level at either end.
Is that the case in your setup? A sketch would be helpful.

 

[256bits]

You might have lucked out with the size of air hole you made.
Probably a pin prick size.
Some great discoveries are made this way, and consider yourself a member of full standing with all the rights and privileges.

Any bigger of an air hole and it might not have 'siphoned'.
Too small and the tank would have drained to a much lower level.

Summary:: Quite by chance I discovered this new (to me at least) way to use a siphon.
I think I know roughly how it works but what do you think?

I guess it's the venturi effect that causes the air to be sucked

I wouldn't call that a venturi effect, as it's not the flow itself that causes the necessary difference in pressure here.. It is just the difference between the atmosphere pressure and the pressure within the pipe due to the small increase in height of the 'sucking' part of the tube from the water level. Otherwise you wouldn't have the regulation of water level that you speak of. ( there is some venturi effect but the flow flow seems too slow to be of concern ) .

 

[fxfocus]

Hopefully the following will make things clearer...

The version in the video includes a small 3D printed part containing the hole.
Dimensions are OD 5mm ID 2mm small hole in side approximately 0.5mm.

You can see this printed part in the video. It is blue and connects the black tube below to the clear tube above. You can also see where the black tube enters the water. The clear tube goes over the side of the tank down to the floor and makes it's way to a drain. Assuming I start from a state of equilibrium I find moving the hole up or down some arbitrary distance will after a short time cause the water level to go up or down by a similar amount.

 

[fxfocus]

It is just the difference between the atmosphere pressure and the pressure within the pipe due to the small increase in height of the 'sucking' part of the tube from the water level. Otherwise you wouldn't have the regulation of water level that you speak of.

So if I read you right... The pressure inside the pipe reduces as you move up the pipe. Somewhere between the waterline and the apex of the pipe there will be a sweet spot at which air will begin to be sucked in. So if the water level rises the sweet spot will rise and vice versa? And am I right in thinking the pressure inside the pipe will decrease when the flow speeds up and vice versa?

 

[fxfocus]

Just to illustrate why I got excited when I found this worked, the following link leads to a hell of a lot of videos offering ways to achieve the same effect but none coming close in terms of simplicity...
https://www.youtube.com/results?search_query=diy+overflow+siphon
I'm not expecting anyone to watch the videos but just the sheer number of them show the level of interest in this kind of application..!

 

[256bits]

So if I read you right... The pressure inside the pipe reduces as you move up the pipe. Somewhere between the waterline and the apex of the pipe there will be a sweet spot at which air will begin to be sucked in. So if the water level rises the sweet spot will rise and vice versa? And am I right in thinking the pressure inside the pipe will decrease when the flow speeds up and vice versa?

Air should be 'sucked' in any place just above the water level.
It is the pressure from the atmosphere that pushes the air into the pipe.
The water surface tension ( at the hole ) has to be overcome, so some small height 'adjustment' there to be considered.
With a full pipe, the water will have a particular flow through the siphon and the water level will fall,
As the water drops some small distance below the hole, air can now enter, so there is now two sources of fluid able to enter the pipe - the water and the air.
As the water level drops further, more and more air can enter, and less and less water.
At some point, with dropping tank water level, the siphon would break break.
Since there is a continuous flow into the tank, the water level is regulated.

I would have thought a smaller hole than 0.5 mm would have been necessary for this to work.

This would be a great product to promote for tank level self regulation.
I don't think everyone who tries this will have success, so you might have a good product there.

 

[fxfocus]

Air should be 'sucked' in any place just above the water level.
It is the pressure from the atmosphere that pushes the air into the pipe.
The water surface tension ( at the hole ) has to be overcome, so some small height 'adjustment' there to be considered.
With a full pipe, the water will have a particular flow through the siphon and the water level will fall,
As the water drops some small distance below the hole, air can now enter, so there is now two sources of fluid able to enter the pipe - the water and the air.
As the water level drops further, more and more air can enter, and less and less water.
At some point, with dropping tank water level, the siphon would break break.
Since there is a continuous flow into the tank, the water level is regulated.

I would have thought a smaller hole than 0.5 mm would have been necessary for this to work.

This would be a great product to promote for tank level self regulation.
I don't think everyone who tries this will have success, so you might have a good product there.

I just checked it out and it is exactly as you say... Air is sucked in the instant the hole rises above the water line. At equilibrium the hole is approximately 2.5cm above the waterline using the present setup.
The hole may well be smaller in practice as my 3D printer doesn't necessarily make a nice clean hole at this scale. Also the hole is slanted down so as to direct any squirting (when the pump is running) down towards the water.

 

[fxfocus]

I think it's worth noting another major plus for this setup is that if the level falls a short distance below the setpoint (maybe because the inflow is interrupted) the siphon breaks, so no danger of emptying the tank.

 

[256bits]

I just checked it out and it is exactly as you say... Air is sucked in the instant the hole rises above the water line. At equilibrium the hole is approximately 2.5cm above the waterline using the present setup.
The hole may well be smaller in practice as my 3D printer doesn't necessarily make a nice clean hole at this scale. Also the hole is slanted down so as to direct any squirting (when the pump is running) down towards the water.

If it works, it works.

Question: Does the the air enter as a continuous stream, or is it broken off as bubbles at the hole?

 

[fxfocus]

If it works, it works.

Question: Does the the air enter as a continuous stream, or is it broken off as bubbles at the hole?

Appears to me to be broken off as bubbles at the hole. If you freeze frame the video you can see the exact makeup and the individual bubbles.

There also appears to be a minimum flow required before air is sucked in. If I run the siphon near the bottom of it's range just before the siphon might break, it goes into a sort of fast slow cycle... As more bubbles enter the tube the flow quickly slows to almost stationary at which point air ceases to be sucked in and only water is entering the tube. As more water enters the tube and bubbles are exiting at the other end the flow begins to accelerate slowly at first but (I suspect) exponentially until air is again sucked in and the cycle repeats. Fascinating to watch for a nerd like me lol.

 

[256bits]

This is fascinating indeed.

At times I would take a hose and attach it to a faucet to fill a bucket. ( One of those rubber pieces on the end of a hose that fits over the tap )
The surging at certain flow speeds ( with the tap more open ) was apparent, as well as entrained air coming in from some place. Never could figure out why that was happening.
Perhaps the 'venturi' effect from a moving fluid is more predominant than what I stated.

 

[Lnewqban]

The column of water-bubbles mix that is formed downstream of the air intake hole weights less (reducing its siphon effect) than when it was formed by water only or by a mix of water and small bubbles.
It is similar principle to the pumping effect achieved in stack tubes of bottom under-gravel water filters.

 

[fxfocus]

The column of water-bubbles mix that is formed downstream of the air intake hole weights less (reducing its siphon effect) than when it was formed by water only or by a mix of water and small bubbles.

I'm sure that's right... The following video illustrates this but also I think points to the venturi having some effect so a combination of the two I think?

Siphon in what I'm calling "cycling mode"

Feel like I'm starting to understand this now as a result of this discussion...

Seems like air is only sucked in above some critical flow rate.
When the flow is at it's slowest the weight in the drop is still just enough to suck water slowly over the apex but no air at this point. With no air coming in the siphon begins to speed up and the next cycle begins...

 

[fxfocus]

Oh and just for fun has anyone any ideas for what to call the little blue plastic thingy I printed that makes the whole thing work?

 

[Lnewqban]

Oh and just for fun has anyone any ideas for what to call the little blue plastic thingy I printed that makes the whole thing work?

Something like this, working in reverse:
https://en.m.wikipedia.org/wiki/Vacuum_ejector