I've been reading about these for awhile, but have only seen a few pictures.
Summarizing what I've read: Use multiple tubes as small as you can...usually at least 50 feet each.
Run the infeed tubes outside of the chamber.
Allow a way to flush the tubes as the aerobic bacteria seems to release a slime coat over time which blocks up these tubes.
Reoxygenate the water before it comes out.
Also, from my understanding, you can obviously smell the hydrogen sulfide if it is being produced, which means your flow rate is too slow.
If you detect nitrites on the output, your flow rate is too fast.
Anytime you open up the chamber, it takes 6 weeks to recycle.
Since the flow rate is slow, you can gravity feed from an overflow and use valves to slow the overall rate to a drip.
Keep the output separate from anything else so you can test the output water.
Keep in mind, I've never built one, this is all the theory I've read.
If you build one, take pics and keep a log of your findings and share them!