It's only taken a couple of years, but I've finally made some fittings for my verical boiler (to the design in Tubal Cains Simple Steam Engines 2, but a little bit taller). My pressure gauge is one of a fire extinguisher, modified by removing a ring which increased its resitance – recalibrated against a quality gauge it works well.

Anyway, I filled it yup, turned on the gas and started heating. At 10psi the gauge leaked wisps of steam around teh base, a quarter turn stopped that. At 15 psi safety valve started to leak a little steam. I had it set fairly loose, and I was concerned that by 20psi it still hadn't popped.

I gingerly prodded the top pin of the safety valve to see if it was stuck and 'wham' a huge jet of wet steam hit the ceiling of my workshop, presumably as it primed. I lost 1/4" of water in the glass but it stopped wjhen I turned the gas off. A repeat outside saw the same., except teh geyser was about 12 feet high and created a localised shower downwind, much to the cat's alarm. It didn't close again until the pressure was nearly zero.

It's a Martin Evans pattern safety valve, and I think the problem is the fit of the valve stem in the screwed and vented plug at the top of the valve, although it has a good polished finish.

At least I know the valve can vent plenty of steam, but any advice on getting a proper 'pop' action?

Thanks

Neil

Looking at Gordon Smth's designs (thanks Mick) there are four very noticeable differences to 'standard' designs'

1. Instead of a small hollow for the ball on the end of the spindle, there is a deep drilling that effectively shrouds the ball creating a shape like a hemisphere on the end of a cylinder. This centralises the ball more positively.
2. The narrow annulus around the ball is much longer and combined with the larger 'socket' means that there is a more constant sized annualr gap above teh ball, so resistance steam flow varies much less with valve lift.
3. Valve lift is reduced (the tall valve is essentailly just the small one at the top of a long base) so teh amount of valve lift is very limited.
4. The spring is FAR shorter so regardless of the spring strength, its rate will be much higher -in fact the springs illustrated probably lock solid when most valves would still have lots of easy travel.

I imagine that these have several effects:

1. The spung ball ensures, like most valves, opening is progressive, but changes have been implemented to maintain this progression is smooth as the valvecontinues to rise.
2. The size of the annulus around the ball is critical, and must be large enough to pass all the steam the boiler can make, as unlike other designs there is no possibility of the valve lifting right up, as when I caused my boiler to prime.
3. For the same reason, as long as the annulus is well proportioned, lifting of the valve is less likely to cause priming.
4. The constant annulus design means valve lift has a smooth relationship to the rate of steam loss – the lack of a 'step' means there is the minimum of hysterisis and stopping the valve lingering open.
5. Finally, combined with the lack opf hysterisis, the high spring rates help ensure a snap action with prmpt closure.

This is what ensures the valve opens and closes again over a small pressure range with of minimum of hysterisis

Neil