Tuesday, March 6, 2012

Bell Siphon - My Conclusion

UPDATE -After 6 months my bell siphon failed to start siphoning.
Here you will find the best way to break the siphon, but I have abolished any further use of Bell Siphons because they are and always will be prone to failure.  
I am now using

Drip Dry Flood and Drain

But if you insist upon a bell siphon the instructions below are your best bet

I have found the answer to the Bell Siphon Blues. I have found a way to make any Bell Siphon absolutely 100% trouble free. Below you can read about my anguish and how I labored to get my Bell Siphon to work. Twice I thought I had it right.  Each time it worked for over a week without flaw, and then quit working.

There is a lack of real engineering available, and even the College of Hawaii failed to provide a mathematical reason for how a Bell Syphon works.   I will not provide the math either, but when you see my solution it will be so obvious that you will not need anything else.

First Stand pipe with 1-1/4 extension.
When screwed in it stands 8-7/8" tall including the bulkhead. 

New stand pipe goes directly to 1" from 2". 
The top is slightly smaller as the first standpipe used a coupling.

This is the tail piece without the 1" male adapter as shown below 

Exactly the same as  above except the horizontal section is 1" longer
It's difficult to see in this photo due to camera angles,
but there is 3/4" of air space above the standpipe,
the bulkhead adds about 3/8" to the height of the standpipe .
I don't believe the air space is critical.  I used to have about 4" of air space and it still worked.  I think the important thing about the air space is not to restrict it by too much.

Here are the differences 

The Tailpiece
By shortening the horizontal section of the tailpiece I shaved 1 or 2 seconds off the siphon break.
By lengthening the horizontal section by 1" the siphon will not break.

The Standpipe
With the 1-1/4" section of the standpipe
  • Drain Time (Siphon catch to catching air) - 3:52
  • Catching air to flow stop     - 0:12
  • Flow stop to burp               - 0:17
Without the 1-1/4 section of the standpipe
  • Drain Time (Siphon catch to catching air)  - 2:45
  • Catching air to flow stop      - 0:17
  • Flow stop to burp                 -0:05

My pump is rated at 500 GPH at 1 foot head with a maximum head of 17 feet.
I have a 2-1/2 to 3 foot head.  It varies due to the level of the sump, but the fish tank buffers any difference in flow.


My preference is the standpipe with the 1-1/4" extension. I ran the tests several times and recorded the events in order to get a verifiable test.  Without a doubt the standpipe with the 1-1/4" extension not only broke siphon faster, but it did it with a more profound and determined manner.

I spent well over 30 hours fine tuning this siphon.  My advice is to find someone who is willing to share their GPH flow and exact siphon design and then follow that.  In the real world your system is going to differ slightly from every other system so even if you try to duplicate another siphon you will probably have to spend time fine tuning it.  Have patience, experiment till you find something that works for you.  Taking a video of the test will prove to be very helpful. 
In my initial testing I used a hose between the sump and the tank.  After hard plumbing the pump into the system my flow rate increased slightly and the siphon no longer broke siphon.  Minute changes can make all the difference. 

Click Here To Watch Test 1
Click Here To Watch Test 2 

I have not tested this, but I believe it best to build your bell siphon so that it can break siphon at a flow rate higher than your system runs at.  If it will break siphon at a high flow rate then it should break at a slower rate too.

Here is a link to Construction of Automatic Bell Siphon from the College of Hawai


  1. Hey Bob,

    Thanks for the videos and insight into the bell siphon construction/function. I initially struggled with the creation of mine. Something that i found was that the length of the pipe below the siphon had an effect on the siphon catching and breaking as well.

    I have my own specific setup that has another container with lava rock, below the grow bed. I initially did not have a tube extending into this container, just the open hole from the siphon tube above (1.5 inch pvc). The siphon was not catching very well, and it was almost never breaking. I modified my bell to have a tube coming from the top, which helped break the siphon, but it was still having trouble catching.

    So, I extended the tube below the container and the siphon worked a lot better.

    I wish I could find a documented case between the Flow Rate (gph) versus Siphon Diameter versus Bell Design.


  2. Having built and shipped just about 1000 bell siphons now (as of 11/20/15) I just wanted to say that if your siphon is not working it just might be your not understanding what makes a good bell siphon design. There is more going on in there than most folks would believe. Nonetheless it ain't rocket science. Ours always work and over a wide range of flow rates and and all media depths from 4" to 12". They work great year in and year out and we have them installed in over 30 countries now. Just say'n. If you need help with yours just join me on FB under Smoky Mountain Aquaponics. If I can't get yours going I'll help you build our design which always works. Sometimes it is just easier to start over and we aren't talking big $ or very much time. Ours average 6-10 minutes to build, look great and work every time.