[Elecraft] PA Transistors Maintenance in K2

[Clay Autery]

It is possible that an explicit torque has never been calculated.  It is 
not a particularly hard thing to do however.

I do not have a K2, so I can't do it, but here's the idea:

1) Fastener size/type, material, thread spec: (e.g. 4-40 x ___ pan head, 
phillips, zinc coated, non-rated steel
2) What does fastener anchor in?  (e.g. aluminum heat sink, what alloy 
aluminum, thickness of threaded area.
3) Thread spec... not JUST the #4, 40 tpi, but the rating for thread 
engagement.
4) Check the specs for the RF transistor.... package, et al.  Docs 
may/should have a max torque spec for the package.... maybe... package 
material, etc.
5) # of fasteners... usually 1 maybe 2.

There are tables that specify general torque specs for fasteners, 
threads in various materials, etc.... The idea is to take the lowest max 
value from the list.  This becomes the MAX NTE.
Then, use some analysis and judgement in what forces are to be imparted 
to the joint....  static tension, sheer, axial, radial, vibration, et al...
Then in this instance consider the qualities of the mating surfaces....  
are the flat?  What is the run out?  Are they smooth? To what grit spec 
are they smooth?  There's a difference between flat and smooth.
What is the interface material being used?  Thermal pad? Insulating?  
Non-insulating? Initial thickness? Read the spec sheet for the thermal 
pad/compound.

I suspect here that the RF package has an insulating thermal pad between 
it and the heatsink...  It was probably spec'd to keep the RF package 
within its operating envelope under 90% of expected operating 
conditions... knowing Elecraft, probably even a higher spec, and maybe a 
little extra to cover the "cheater".

Thermal pads and MANY top quality thermal compounds REQUIRE a curing 
period.  If the package is held with a tension spring/fastener, there is 
usually no requirement/need to re-torque.

IF it is JUST a screw and maybe a locknut, you NEED TO RETORQUE after a 
given number of thermal cycles....  The thermal interface material flows 
under heat and cools when not in use.  Eventually it reaches an equilibrium.
You torque the fastener JUST enough to hold the junction static at this 
point.  I'm a freak, I would use a thermally conducting thread 
locking/anti-seize compound (I made one).
I NEVER torque unlubricated fasteners.... ESPECIALLY steel to aluminum 
interfaces.

The final torque for this application is likely in the relatively low 
inch-pound range.... just guessing.  I have 2 inch-pound wrenches... 
0-15, and 0-70.  Most people don't....

Key... do NOT over-torque.  Consider the lever length and force applied...

73,
Clay, KY5G


On 10/29/2017 8:03 AM, David Woolley wrote:
> As far as I can tell, no torque settings are given for any of the 
> fasteners, wouldn't it be an idea to provide those, for people who do 
> have suitable tools.  Generally there are just warnings not to 
> over-tighten.
>
> On 28/10/17 20:24, Mel Farrer wrote:
>>   If the people at Elecraft have done their job and mounted the 
>> transistor properly with the correct torque on the mounting hardware 
>> procedure, no additional maintainance is required.
>
> -- 
> David Woolley
> K2 06123