[AR] Re: Updated custom ball valve

  • From: Peter Fairbrother <zenadsl6186@xxxxxxxxx>
  • To: arocket@xxxxxxxxxxxxx
  • Date: Tue, 24 May 2016 19:32:31 +0100

On 21/05/16 23:57, Robert Watzlavick wrote:

I was never really satisfied with my semi-custom cryo ball valves (see
aRocket archives around 2014-11-24).  These valves use the ball and body
seals from a Swagelok SS-62T6 but with a new aluminum body, stem, and
stem seals.  I had some stem leakage due to shaft wobble that I was able
to address by adding a 2nd spring energized PTFE seal and pressurizing
the seal cavity but I had the feeling I was approaching it the wrong
way.  Several folks made good suggestions including Peter who provided
some sketches of alternate configurations.  I had wanted to get the
bearing closer to the seal so I spent some time designing and
fabricating an alternate arrangement the past few weeks.  One thing I
discovered is that even with two bearings on the shaft, there is still
an excessive amount of free play until the bearings have some axial
preload. Apparently, bearings (at least standard ones) have to have a
minimum amount of preload on them just to meet their specifications.  I
set up a test fixture with two bearings separated by about 0.5 inches
and even with the outer races held tight in a collet, I could wiggle the
shaft up to about 0.005 inches by pushing sideways with my finger. After
some of reading about bearing design, I modified the valve body and stem
to use 0.0005 inch slip fits and then added two sleeves between the
bearings.  The preload is applied to the inner races through an inner
sleeve (recommended for maximum stiffness) and the outer sleeve is used
to help keep the bearings aligned with each other and to press them into
the housing without damage.  Here are some sketches of the new design:

http://www.watzlavick.com/robert/rocket/rocket1/drawings/ball_valve_3_assy-annotated.pdf

The inner sleeve prevents the preload nut from doing anything useful. It just tightens the distance between the bearings - which you don't really care much about.

Tightening the nut actually _lessens_ the force on the seal ... and increases the useless torque needed to turn two unnecessarily-squashed-together bearings.



I think you need to ignore the top bearing, and just worry about the lower bearing, which should be pressed against the seal with a belleville washer. You will need some kind of screw adjustment on the body (tap the body, don't thread the stem) to do that.

sorry no, that's not clear.

you have two seperate issues, the force on the stem and the force on the seal. There is only one adjustment needed, the force on the seal.

The top bearing can just sit there, floating up-and-down on the stem, with no up-and-down force on it, only concerned with keeping radial position.

The lower bearing should take the up-and-down forces on the stem. Within reason, it doesn't need to be adjustable in position to do that. upwards motion can be mechanically constrained, and downwards motion can be constrained by the seal.


so, from the top we have: top bearing (floating), space, nut screwed into body with shaft running through it. bearing between nut and washer, belleville washer, seal. The bottom bearing should perhaps be angular contact.



The seal is still too far from the ball.

- Peter Fairbrother

(in a hurry, will try and look closer this weekend - I think I can do better than this:



Attachment: rect3264.png
Description: PNG image

Other related posts:

  1. Home
  2. arocket
  3. Archive
  4. 05-2016