So, we had a problem… the flame was shooting into the pipe. I ran many experiments. I lengthened the accelerator pipe holding the jet, I pushed the jet further into the tube, I widened the air-intake channels, all to no avail.
That’s when I was left with 2 options… the jet needs to be smaller to get faster propane velocity or the regulator is broken. I’d noticed at one point that I could no longer tune the jet with the regulator as I had in my initial firing.
That’s when I realized that maybe it wasn’t the regulator, maybe it was the propane tank! A little googling later…
A second cause might be the automatic activation of a surge protection device within the propane regulator. If, for example, you turn on the tank valve before you fully turn off each of the burner knobs on the grill the surge protector will likely sense a leak and activate. The fuel flow will be very low. The remedy is to turn everything off, disconnect the tank, and reconnect everything before starting over.
I disconnected everything, turned on the tank and started dialing up the regulator to mid-pressure (15 psi or so?) before trying to light it… the gas was moving so fast that I couldn’t even light the torch. Then I started decreasing pressure, attempting to light, decrease, attempt, etc. Eventually the torch lit up… and went back into the tube!
So, now I know that the burnback is strictly a function of low propane flow, not an intrinsic problem in the burner design. The real problem is that the exit velocity is too high at the pressures that prevent burnback. This means that the flare isn’t doing its job… the gas hasn’t slowed down enough at any point along the flare for the combustion speed to equal the fuel speed, which results in a stable flame. This might be due to an insufficient flare length to get to the proper width that would balance things out.
So, one remedy would be to widen the flare to hope to hit the sweet spot. Or, we can add turbulence at the end to create a “flame holder”, “flame retainer”, or “flame stabilizer.” There doesn’t seem to be any authoritative way to create a flame holder. As there are an infinite number of ways to generate turbulence, that probably shouldn’t surprise anyone.
The simplest solution I’ve heard is to use a hacksaw to carve slots perpendicular to the pipe and slightly taper the end. That taper may have been my main problem… I removed the natural taper from my burner tube in initial fabrication as it seemed like a bad idea to constrict the flow. Now I know better!
Other plans I’ve seen involve a number of smaller, slightly interfering jets formed into a sleet block which is pretty much just a variant on a grid placed over the burner exit.
Putting a cone or other shape down the center of the burner output is described as very reliable… gets the burner to curl around the obstruction and slow down.
In the interest of keeping it replaceable, I think I’m going to be modifying my castable flare to include some turbulence to slow thing down further. What variety of flame holder remains to be seen. I think I may add a cone held in place by 3 fins. Maybe even give the fins a small swirl.
This governmental web page shows the different flame characteristics that result from different flame holders.
I need to think about it.
Here’s a page about designing RAMJETS that discusses types of flame holders.
Update: Saw a youtube video of a guy building his own afterburner using a Fan Guard. Concentric rings design… check. Premade. Check. Wires to attach to the burner… check. I’m going to give this a shot. My little burner may not need the extra metal-flares though. I’ll try it just stock first.