Thanks, Ralph, for posting. This is a reoccuring problem with tailwheels.
That site is a commercial site hosted by an EAA member. EAA has some fantastic maintenance expertise within their community. But keep in mind they are largely concerned with advising owner/builders of
experimental aircraft, therefore they frequently do not make the distinction between altering an experimental versus a certificated aircraft.
While the article mentions a J-3, it suggests a couple of unapproved methods of correcting shimmy. One suggestion is altering the assembly by introduction of a non-approved part, a "pad" beneath the mainspring to alter the angle of the "kingpin". The other suggestion is to bend the mainspring at an automotive/spring shop. The latter suggestion does not address the issue of unduly/permanently stressing the mainspring without proper data. Neither method is approved for certificated aircraft.
To it's credit, the article mentions the "kingpin" angle or caster. This has been addressed several times here at our forums.
The correct caster or angle of the "kingpin" (the shaft/bolt around which the tailwheel actually pivots) is clearly illustrated in the Scott assy drawings found in several places including The 170 Book, but I really like the mentioned article's illustration. (Thanks again, Ralph, for posting the reference.) The top of the "kingpin" should be aft of the bottom.
CAUTION: THIS IS NOT A CESSNA 170 INSTALLATION in the picture. Ignore the tailwheel mainspring shape/curve/geometry. This seems to be a Piper product installation. BUT THE SCOTT 3200 ANGLES AND SETUP ARE THE SAME AS FOR A 170 and the tailwheel should be at similar caster angles. YES! This will cause the tailwheel to lay over slightly on it's side during ground handling when the tailwheel is deflected 90 degrees.
Other message posts regarding shimmy can be found here at:
http://cessna170.org/phpBB2/viewtopic.php?p=2734#2734
http://cessna170.org/phpBB2/viewtopic.php?p=7580#7580