Author Options:

How to turn on and off a stream of gas? Answered

The exact setup I have is a little difficult to explain, but basically I'm designing an extremely simple heat engine, based on boiling ethanol, and need to be able to mechanically alternate the flow of vapour.

I have a little boiler, sealed except for a pipe out of which the vapour escapes. This lifts the column of ethanol in the pipe, pouring it into a water wheel which turns. The ethanol then runs back through a second pipe into the system to be pumped up again.

It works, and fairly well since I'm not looking to get a lot of power or efficiency, but only if I pinch the tube on and off. This allows the bubbles to collapse, which sucks in more ethanol to replace them. When I release the tube the bubbles rush out, lifting the eths, and repeat. I have two valves to keep the system one directional.

So now I'm looking for a way to automate the opening and closing of the gas feed. I've been trying a few things but nothing seems to work well enough yet.

The other consideration is that; due to the nature of the larger project, whatever I use has to be extremely simple to construct, preferably from reclaimable materials.
1 second on, 1 second off seems to work best.

Any ideas?




Something I'm thinking (not sure if this makes sense)


(why can't instructables embed animated gif without them going crazy?)

The slidey thing is a cut in half syringe. The rest is just two springs, a ball bearing, two bolts, one nut and some sellotape.

(Plus pipes, a cork, and a T junction.)

This is to provide a threshold on the opening and closing of the valve, it has to build up a pit of pressure, then opens fully, then the pressure has to drop a bit before it closes again.

I've been thinking of this for the last few days but couldn't think of a way to seal the cylinder til I thought of using a syringe.


Other ways of doing it?

I think you might be better with the cold-feed lower down?



It's a bit hard to explain this without an animation.

1. Vapour emits from the boiler, raising the liquid column. The valve on the feed from the reservoir ensures all that fluid goes up the top pipe.

2. The gas feed is cut off and the vapour collapses. This sucks liquid in from the reservoir below the wheel. The valve on the top pipe ensures the lifted fluid doesn't slip back.

3. The gas feed switches back on. Goto 1.

This repeats until the device has caught up with the sun and the boiling shuts down. The remaining vapour in the boiler collapses and sucks liquid into the boiler, refilling it. The boiler has to cover a maximum of 25º rotation, at up to 100 ml per degree = 2.5 l. The boiler holds about 15 ml eths x 450 gas expansion ratio = 6.75 l.

This isn't a geyser pump, tho that was what I was originally trying to design. It's really just a simple piston, except the piston is the liquid - gas boundary rather than a solid rod.
Push out - suck in - repeat.

It works a lot better than a geyser pump, especially with the small amount of energy I have available.

Pinching increases pressure in the system, and raises the boiling-point. When you release you get a flash-boil by lowering the boiling-point quickly.
More heat is your best answer, after that - less volume/liquid/narrower-bore. You want this to run continuously without the pinch I think.


The boiling isn't the problem, I can fill the whole pipe with vapour, and that  does actually pump a bit of the liquid, but not really enough.

What closing and opening the flow does is turn the vapour / liquid boundary into a kind of piston by proxy. With a constant vapour stream that boundary doesn't move much, so there's not a lot of pumping action. But when the feed is cut off the bubbles collapse, sucking in the liquid, like a piston withdrawing. When it's opened the bubbles force the liquid back again, like a piston extending.

I've tried all kinds of different permutations on this, and this setup is by far the most effective. If I can just get this on off thing sorted...

I seem to remember mentioning coffee-machines. You're giving me the vision of a vapour-shunt, where the system is spitting-liquid under vapour-pressure.
If that is the case, get a cheap coffee-machine and observe how that functions. Then utilise the non-return-valve from it.


I think they work through blasting shedloads of steam through, which kind of sucks in some liquid as well. I'm looking to get a bit more liquid with a lot less vapour. I can insulate the pipe so the bubbles don't collapse at all, and that does work to an extent. But without a downcycle to draw liquid into the line, I pretty much just get bubble at the other end.

I've been using little removable valves from bike inner tubes. They work pretty good. Also playing round with bearings and washers, which don't seal quite as well but are still fairly useable.

Ball-valves should be OK, but maybe use plastic instead of metal. The coffee principle is to generate vapour at the bottom, which pushes liquid ahead of it further up. Then the vapour cools, contracts and the heater re-fills from the cold side of the system. I've had a stab at a ball-valve diagram



Ok, so had another look at how coffee makers work.
Their dynamic is a bit different to mine, in that they blow out the water which generates the steam with the steam. I don't really want to do this as it takes about 2-3 minutes each time to get my ethanol boiling, and there's only 20 ml in the boiler. I want to do up to 25 corrections at 50 - 100 ml per correction.

So I need to isolate off the 20 ml of ethanol and utilise all the vapour it generates to pump the rest of the eths. Therefore the flow of vapour needs to be switched on and off, to mimic the 'all the water in - all the water out' thing that the coffee maker does.

Yes, in terms of power and efficiency. But the only thing I need really to get out of this is extreme ease of construction.
I only need to lift about 100 ml of eths 30 cm in 1 minute, which minus this one last problem it seems to do rather nicely.

.  A relief valve may work. Or a check valve.

Should've also mentioned; no electrics / electronics / motors etc.
The only available energy source is the boiling ethanol itself, heat and pressure. Not a huge amount of either.
But I'm looking to get about .4 Newtons out of 60 Watts, so efficiency and power aren't really issues.

if you're looking to recreate exactly what you're doing now manually (i.e. pinching a tube...i'm assuming some kind of plastic) on a larger scale (or even the same scale) automatically, maybe you could use an electric motor, that's geared to make two revolutions a second. then attach a wheel on the end of the motor that has some form of protrusion. this protrusion would push on a dowel or a stick or something that would squish the tube between the stick and another stick, then the wheel would keep rotating around. it would take roughly a second for it to pass through the beginning and end of the squish, and roughly another second to make it back around to start the squish again.

If you want to automate, you need something like a solenoid valve.  For a name-brand item, check Swagelok's Web site, or search McMaster-Carr's catalog.  Once you've identify the technical name for what you want, then you can do a Google search with that name and find something cheaper.