Screws and Gaskets

[drawoh]

I have just written instructions for taking apart and re-assembling an optical window mount, consisting of a housing, gasket, window and clamp. The window sits on the gasket. The clamp is screwed down tight to the housing, clamping the window and leaving the gasket at a nominal thickness. I specified thread-locker, and a screw torque.

I have just been asked what happens if the person working on this thing does not have a torque wrench.

I have worked out that the screw can be snugged down, and rotated 1/16 turn without exceeding its yield strength. Unfortunately, this only works if there is no gasket. I may just have to say something like "torque it down hard".

Has anyone else had this problem? Any ideas?

--
JHG

 

[MikeHalloran]

Put a pocket/depression/rabbet in the clamp, so that the part of the clamp with the screws seats on the substrate without a gasket, making the torque spec 'tight'.
Control the gasket compression with tolerances on the window thickness, the gasket thickness, and the pocket depth.

OR,

If that's not possible, look at the accessory rocker cover spreaders that used to be sold for small block Chevy engines, comprising a washer with two wings, arched so that the wings flatten as the screws are tightened. The most reliable way to get a specified clamp force is to design the washers so they are still arched just a little at the design force, then instruct the installers to tighten the screws until the wings flatten, then back off half a turn.


<sort of related>
I encountered a ceramic slide valve that was held together by a double ended stud that retained a stiff compression spring that provided consistent force holding the slide valve parts together. ... but it took half an hour of tweaking of the nuts on the studs by an experienced assembler to get the force 'just right', whatever that was, by turning the nuts and sliding the valve parts.

I made some measurements of assembled valves and figured out what force was 'just right', and selected a spring with a much lower rate, selected so that half a turn off coil-bind was just right. So the assembler instructions devolved to 'tighten both nuts until the spring coils close and it becomes difficult to tighten further, then back off one nut, not both, one quarter to three quarters of a turn.

Reducing the adjustment time from tens of minutes to a few seconds, and making misadjustment far less likely.

</>



Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

It is always to your benefit if you provide as much detail as possible regarding your situation. The installation you describe using screws, a flange and a clamped gasket to seal/secure an optical window is not a simple thing. Factors such as operating temperature & pressure, CTE mismatch, etc can make it difficult to ensure a reliable seal with optical windows.

How many screws are you using to retain the window? What size are the screws? What are the dimensions of the fastener pattern? What is the gasket material? What materials are used for the housing, screws, retainer and window? What temperatures and pressures is the optical window subject to?

When multiple screws are used to clamp a flat gasket, the torque sequence can be more important than the final torque. As for using Loctite to secure screws, make sure it can handle the operating temperatures. In order for Loctite to perform properly, it requires all the thread surfaces to be very clean. Making sure the threaded part surfaces are sufficiently clean can be a problem when reassembling used components. And there is no way to verify a threaded connection using Loctite is properly secured.

 

[dhengr]

Drawoh:
Turn the screws down finger tight (snug, just touching the clamp?), then turn them “x +/-” more turns. How far does the screw advance per turn? Does one turn compress the gasket sufficiently for your purposes, or does it take 2.5 to 3 turns?

 

[MikeHalloran]

Additional thought:
This is one instance where I might consider specifying helical split lockwashers.
... not to replace the loctite, but to act as force indicators.
Just tighten the screws until the lockwashers go sorta flat.
Narrow series lockwashers under Fillister head screws are most appropriate for this sort of thing, given their squarish cross section, and OD that leaves the splits mostly visible.



Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

Consider what was stated in the OP. It was said that the margin between no preload and yielding of the screw was 1/16th of a turn (22.5deg). If the screw thread is say 10-32 UNF, that 1/16th of a turn would amount to just under .002" of axial displacement. Not much room for error.

 

[MikeHalloran]

The 1/16 turn | .002" is just a measure of the joint stiffness without a gasket.
We here have no idea what the stiffness with a gasket is,
and we have no idea what the desired range of bolt load is, or if it's reasonably achievable.

Almost flattening a lockwasher is still pretty crude, but it's easier and requires less instrumentation than specifying bolt torque or stretch. Specifying go/not go feeler gages above the lockwasher opposite the split is about as far/insane as I'd go.






Mike Halloran
Pembroke Pines, FL, USA

 

[MikeHalloran]

Additional thought:
I could swear that I've seen hex headed bolts with a second head separated by a deep grooved feature, engineered such that the distal head twists off at some predetermined torque, leaving an ordinary looking bolt with a nub on what would ordinarily be its marking face of the formerly proximal, now only, head.

I couldn't find them, but I did find some generally similar bolts that are not wrenchable once the distal head twists off:

http://www.losspreventionfasteners.com/store/products/tork-bolts/

Mike Halloran
Pembroke Pines, FL, USA

 

[MintJulep]

A Belleville spring or "Load Washer" matched to the gasket stiffness.

http://www.schnorr.com/products.html

Tighten until flat.

 

[drawoh]

Consider what was stated in the OP. It was said that the margin between no preload and yielding of the screw was 1/16th of a turn (22.5deg). If the screw thread is say 10-32 UNF, that 1/16th of a turn would amount to just under .002" of axial displacement. Not much room for error.

I have an M4X0.7 screw clamping 20mm. The screw is A2-70 stainless. When everything is clamped, the screw is clamping two pieces of metal together. I also figured that a half turn of the screw would work fine, if the screw is systematically used once, only. Unfortunately, I am writing a field maintenance procedure, and I cannot assume a torque wrench or replacement screws.

The gasket sits in a pocket, and there is a millimeter of compression of it, controlled by a stack of [&pm;][&nbsp;]tolerances. My original assumption was that the two pieces would assemble solidly before I tighten the screw, and this is wrong.

My primary concern is that they tighten the screws down hard. Sometimes people think "Glass! Don't use force!". The gasket controls the force on the glass.

--
JHG

 

[MikeHalloran]

Well, then, 'twist the shit out of the screws' won't bother the glass.
You probably could control the handle diameter and shape of the screwdriver used, to crudely limit the torque, and specify no wrenching of the screwdriver.


Mike Halloran
Pembroke Pines, FL, USA

 

[MintJulep]

"Tighten until the screw breaks, then back off 1/4 turn."