Thoughts on the design and build of safety valves

In Peter and Matthew Keningtons articles on rebuilding their Romulus (EIM, Dec/ Jan) I have some concerns regarding proportions of the valve parts, operating behaviour-related design issues and material selection for the spring.

The best safety valve designs I have come across are the ones by Gordon Smith, published in the Feb-apr 2001 issues of EIM. I have built several in the series to Mr

Smith’s designs and can confirm they operate beautifully with a gentle opening and positive re-seat action.

One key point is the sizing of the face of the ball plunger end, next to the ball. It is very important to make this accurately, it has a major effect on the harshness/gentleness of the action, because its area is calibrated to the steam flow to provide a balance of force from the steam flow against the spring.

Mr Smith also found the optimal relationship between the ball seat diameter, ball diameter, and ball chamber diameter for good operation.

All of Mr Smith’s designs specify a purchased spring from Lee Spring (www. leespring.co.uk), which will be consistently made and are proven to work for each valve design. I suspect that if built to the specification in the latest articles, the valves will be harsh in operation and may not re-seat well.

I suggest the authors have a look at Mr Smith’s SV#7 design – it looks like a comparable-sized safety valve to the one they describe but the key diameters and ball size are different than the article sizes and this will make a big difference I think to the operation of the valves. If they follow Mr Smith’s design as closely as possible and use the recommended purchased springs, they will not be disappointed, and with no need to reinvent anything from first principles.

Regarding spring material, I suggest that plated steel or unplated music wire should never be used for a safety valve due to the risk of corrosion and possibly wire wastage. This wastage could significantly affect the valve’s reliable operation.

If the plated spring is replaced frequently the risk of corrosion is greatly reduced, but this assumes that owners/operators (maybe several sets of people over the life of a locomotive) will have the knowledge passed along in order to know to change it. If a stainless spring is used it will last many years with minimal if any risk of corrosion significant enough to affect its operation.

One further note – in the article the ball seat is shown as machined directly in line with the edge of the hex on the outside of the valve. It would be much better from a stress perspective to have the seat and hex edge offset one way or the other by some amount, to avoid a stress concentration/stress raiser in the body. The body could see high forces applied to it from accidental side impacts.

If the body ever split at a stress raiser while under steam there would be a big bang and steam release at best, or potential injury to persons close by and boiler damage from rapid depressurisation at worst. This is just food for thought.

Thank you for your comments – it is always good to have feedback on what we produce and to learn from the thoughts of others.

We were aware of Gordon Smith’s articles on safety valves and, if starting from scratch, we would have almost certainly gone down this route. The aim of the article was not, however, a treatise on the optimal design of safety valves but a discussion of a problem we encountered and a means of then rectifying that problem within the envelope of the original safety-valve body.