[ARC5] Arc-5 screw removal

[Mike Hanz]

In my previous post, I wasn't sure what you had in the way of tools, 
Neil.  In those kinds of cases I generally revert to describing a 
consistently successful technique, counting on the reader's ingenuity to 
adapt it to his environment, but it does usually have the need for some 
more sophisticated tools.  Your procedure is correct to a point, with a 
pilot hole and subsequently larger bits.  In the web link I provided, I 
mentioned that you can use a *small* hand drill if you are careful - a 
stiff short bit like the #1 center drill described makes it easier to 
get started.  I did it for years before acquiring the mill, with fair 
success.  The example is for small rivets but the principle is the 
same.  Brian's suggestion of a PC board drill is also excellent, and if 
one has a Dremel tool, is the way that I would approach it.  It will be 
necessary to get the shortest drill you can find - the spiral portion, 
not the shank - most of the ones on ebay seem a bit long for this kind 
of work.  Using a magnifying visor, you can generally adjust the center 
of the hole slightly in the first few revolutions of the drill point 
with a bit of sideways pressure.  If you are off center, you will have 
to be satisfied with a C-shaped "husk" to get out of the tapped hole.  
The major problem is that almost every command set screw that I have 
found broken off was due to corrosion in the threaded area, which makes 
it more recalcitrant than simple mechanical deformation.  There are 
reverse spiral drills that you can buy that may help, but what you are 
suggesting in #3 below is actually a form of easy-out, and the blunted 
drill tip will forced the husk outward, locking it into the corrosion 
even more.  At least that what I've experienced.

Using the gradual enlargement technique, eventually you will reach the 
point where the drill flutes will be just touching on the minor diameter 
of the threads in the insert.  You can do a bit of "sideways milling" 
with the drill by moving it up and down while putting gentle pressure on 
it toward the "thick" side of the husk, but it takes a lot of patience, 
and the chance of snapping the bit with this technique is higher than 
straight drilling.  I would suggest putting away the drill bits at that 
point.  That's when I start trying to tease the husk out of the hole 
with dental instruments.  Brian's source for dental tools is where I 
obtained mine as well, but sadly here in the US your dentist may not be 
cooperative because of concern about lawsuits.  I see them at hamfests, 
though.  If the husk resists all efforts to persuade it from coming out, 
then I'll unlimber the tap, with the realization that there is no 
guarantee that the tap will follow the original threads.  In fact, you 
can almost guarantee it won't.  But at least you will have a semblance 
of a thread in there.  Try it, you'll like it...

Employing Colin Powell's Rule #4, here's the bottom line from my 
perspective:  You don't have to get a perfectly centered hole to be 
halfway successful.  The closer you can do it, the better, but in the 
final analysis, you can run a 3-48 tap into whatever mess you have 
caused and the screw will usually hold "well enuff" unless you try to 
torque it to original specs.  No one will be using this in an aircraft, 
so you don't need to obsess about it.

73,
Mike

On 3/11/2012 11:31 PM, Neil wrote:
> The consensus seems to be that, other than using a dentist's pick
> to try to rotate the broken-off section out of the threads, the
> operation requires precision drilling and is beyond the ability of
> the average guy who has only a cordless drill.
>
> I had been thinking along the following lines:
>
> 1. drill a small pilot hole in the centre of the shaft
> 2. heat or cool (I'm not sure which) the entire item to take advantage
>      of the different coefficients of expansion of aluminium and brass
> 3. drill at very low speed into the pilot hole with a blunted bit whose
>      diameter is somewhat larger than the pilot hole and somewhat less
>      than the minimum thread diameter.
>
> The idea being that the blunted bit would bind when entering the
> pilot hole and slowly rotate the brass screw out.
> I don't have the workshop to carry out the procedure. Does it have
> any merit?