L.E. American & Feedwater?

This forum is dedicated to the Live Steam Hobbyist Community.

Moderators: Harold_V, WJH, cbrew

User avatar
Greg_Lewis
Posts: 1571
Joined: Wed Jan 15, 2003 2:44 pm
Location: Fresno, CA

Re: L.E. American & Feedwater?

Post by Greg_Lewis » Fri Sep 21, 2018 10:17 am

And remember, if you increase the diameter of a circle by the square root of two (1.4142) you double the area. So taking a 1/4-inch ram and bore up to .3535 you should get twice the water out.
Greg Lewis, Prop.
Eyeball Engineering — Home of non-interchangeable parts.
Our motto: "That looks about right."
Celebrating 30 years of turning perfectly good metal into bits of useless scrap.

User avatar
DianneB
Posts: 672
Joined: Tue Aug 21, 2012 3:05 pm
Location: Manitoba, Canada

Re: L.E. American & Feedwater?

Post by DianneB » Sat Sep 22, 2018 7:54 am

I took detailed measurements of the existing pumps and think the basic problem lies with the check valves.

The checks use a 0.1875 SS ball in a 0.240 dia hole. They have a square seat opening into a 0.156 passages. But there is nothing to limit the ball from following the flow - total lift 0.229 - and at that point the ball will plug the discharge hole!

I can't believe that a good designer like Martin Lewis of Little Engines designed it this way, unless there was something to stop the ball from lifting too far, some little piece that has been lost over the years. (Does anyone have drawings for the LE crosshead pumps that show some kind of 'ball stop'?)

I am tempted to pick up some stainless spring and spring-load each ball against its seat - that would also limit the ball trying to "go with the flow" and blocking off the outlet and would also make the check more fast-acting.

Thoughts? Suggestions?

Harold_V
Posts: 17201
Joined: Fri Dec 20, 2002 11:02 pm
Location: Onalaska, WA USA

Re: L.E. American & Feedwater?

Post by Harold_V » Sat Sep 22, 2018 5:34 pm

A spring offers the possibility of stopping (or seriously restricting) flow in the desired direction, much the same way as no spring. An installed obstruction that allows fluids to bypass when the ball is seated in the wrong direction would be far more desirable.

H
Wise people talk because they have something to say. Fools talk because they have to say something.

RET
Posts: 756
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: L.E. American & Feedwater?

Post by RET » Sat Sep 22, 2018 6:36 pm

Hi Dianne,

You may have a number of problems.

First; yes you definitely need to limit the ball travel. If you don't, it takes too much time when the ram reverses for the pump suction to pull the ball back onto the seat. Look at it in terms of ram travel vs ball check travel. If the ball lifts a lot, a significant part of the ram travel occurs before the ball seats, thus that part of the ram travel is "wasted."

If this is the only check between the pump and the boiler, this source of loss may not be quite as significant, because boiler pressure may force the ball onto the seat but you still need to keep the ball travel down because while the ball is "traveling," the pump isn't pumping. Common practice for 3 1/2" gauge is between 1/32" and 3/64ths ball lift. The loss will also be speed dependent.

Harold is right; if the spring comes up "solid," it will either partially or completely block the pump discharge, depending on how the spring seats at the discharge end of the check. A little thought will show you that this is true.

Are there two crosshead pumps or one? It isn't really clear from the posts. Second, are the check valve(s) horizontal or vertical?

The second potential problem is that under the right conditions, with a "hard" seat, the ball can spin on the seat instead of actually seating properly. My way of eliminating this potential problem is to make a "soft" seat by using an "O" ring. This works just fine if you either design it so the "O" ring is "captured" or if you make the "O" ring seat so the flow doesn't tend to lift the ring out of place. You can do this by grinding a special cutter so that the"O" ring sits in a special groove that extends part way into the inside of the ring. In this way, the main liquid flow is directed through the "O" ring without lifting it. I have done this in the Boston & Albany and it works just fine. If the ball spins on the seat, you can hear it. Eventually it seats, but not before a significant portion of the pump stroke has happened.

Sometimes its the little things that trip you up.

Richard Trounce.

User avatar
DianneB
Posts: 672
Joined: Tue Aug 21, 2012 3:05 pm
Location: Manitoba, Canada

Re: L.E. American & Feedwater?

Post by DianneB » Sun Sep 23, 2018 5:31 am

Thanks for the comment Harold. I have considered that possibility.

RET: There is a boiler check in the discharge line so I presume it would seal first because it has a limited ball lift.

There are two crosshead pumps operating 90 degrees apart and all checks are vertical. I had considered adding an o-ring as a 'soft seat' and the possibility of the flow lifting the ring. I'll have to draft it out (in CAD) and see if I can find a ring of appropriate size.

I may have more time to work on this than I had hoped! :( The weather has turned bloody COLD and there is no relief in the forecast.

User avatar
Bill Shields
Posts: 5169
Joined: Fri Dec 21, 2007 4:57 am
Location: Somewhere in the World
Contact:

Re: L.E. American & Feedwater?

Post by Bill Shields » Sun Sep 23, 2018 6:46 am

there is no need to reinvent the wheel here.

a hard seat with the ball lapped to fit the seat and a lift of no more than 10% of the ball diameter is all that you need. a good way to do this is to glue a ball to a piece of brass tubing and use some lapping compound to mate the two together

lift limit should be against some sort of 'relieved' stop that will allow the water to pass while the ball is there - typically cut an X with a saw or make a 'clover leaf' cut out with an end mill.

This recipe has worked for 100s of pumps for better than a century.

User avatar
DianneB
Posts: 672
Joined: Tue Aug 21, 2012 3:05 pm
Location: Manitoba, Canada

Re: L.E. American & Feedwater?

Post by DianneB » Sun Sep 23, 2018 8:17 am

I always look forward to your advice Bill!
Bill Shields wrote:
Sun Sep 23, 2018 6:46 am
a hard seat with the ball lapped to fit the seat and a lift of no more than 10% of the ball diameter is all that you need. a good way to do this is to glue a ball to a piece of brass tubing and use some lapping compound to mate the two together
10% is the number I couldn't find! I drafted the whole thing out in AutoCAD and measured the lift to give the same cross sectional area between the seat and the ball as the area between the ball and the 'downstream' bore. I came up 8% of the ball diameter.

I like the idea of a ball glued to a stick or tube to clean up the seat!
lift limit should be against some sort of 'relieved' stop that will allow the water to pass while the ball is there - typically cut an X with a saw or make a 'clover leaf' cut out with an end mill.[/quote)

That's what I was thinking - a piece of round brass to fit snug in the bore and then mill it into an X.

User avatar
Bill Shields
Posts: 5169
Joined: Fri Dec 21, 2007 4:57 am
Location: Somewhere in the World
Contact:

Re: L.E. American & Feedwater?

Post by Bill Shields » Sun Sep 23, 2018 10:40 am

8 or 10% is not critical...just keep it low so the ball does not float. These numbers are in LBSC stuff from before I was born.

Springs are more trouble than worth since pressure differeentials will take care of it.

User avatar
Greg_Lewis
Posts: 1571
Joined: Wed Jan 15, 2003 2:44 pm
Location: Fresno, CA

Re: L.E. American & Feedwater?

Post by Greg_Lewis » Sun Sep 23, 2018 11:36 am

Be sure to check the sizes of all the spaces in the pump and piping. In my file on pumps I have a printout from an unknown internet source, the author being only “Don.” He writes;
“The optimum diameter for the bore that the ball seats on is .707 x o.d. of ball. This will ensure that the contact angle is 45°.
The minimum diameter of the bore that the ball moves in is 1.31 x o.d. of ball. This will only just give enough clearance for the liquid to flow past the ball. It is recommended that 2 x o.d. of ball is used to provide reasonable flow conditions around the ball. Lift is no more than the diameter of the ball. 1/2 is plenty, 1/3 is nice.”

And, of course, do the math to be sure that there is enough space above the ball for the water to flow with minimal impediment to the outlet.

A Google search of some of the phrases in the above quote brings up this thread on check valves:
https://steamautomobile.com:8443/ForuM/ ... 7108,27205
Greg Lewis, Prop.
Eyeball Engineering — Home of non-interchangeable parts.
Our motto: "That looks about right."
Celebrating 30 years of turning perfectly good metal into bits of useless scrap.

User avatar
Bill Shields
Posts: 5169
Joined: Fri Dec 21, 2007 4:57 am
Location: Somewhere in the World
Contact:

Re: L.E. American & Feedwater?

Post by Bill Shields » Sun Sep 23, 2018 3:36 pm

I think that 2 x ball OD is too much.....will float around and beat the seat to death over time.

I typically like 1/32" hole diameter over ball diameter. It is more than enough

User avatar
DianneB
Posts: 672
Joined: Tue Aug 21, 2012 3:05 pm
Location: Manitoba, Canada

Re: L.E. American & Feedwater?

Post by DianneB » Mon Sep 24, 2018 5:39 am

Greg_Lewis wrote:
Sun Sep 23, 2018 11:36 am
It is recommended that 2 x o.d. of ball is used to provide reasonable flow conditions around the ball.
I too think 2x ball diameter is excessive. At that size, horizontal movement of the ball can become a problem.

The criteria I used to verify the existing check valve was to determine the cross-section area of the supply line and the annular area of the cross-section between the ball and the chamber in which the ball sits. As long as the annular area is greater than the supply line area, it will not provide additional restriction.

Post Reply