US coils stamp are often faked; some say as many as 80% of all Washington/Franklins coils are faked using various techniques. Learning what to look for is key when identifying your stamps.
1. Measure distance between cut edges (horizontal coils are 25mm, vertical are 21.5mm). A variance of + or - .03mm is tolerable. Anything more can be discounted as fake 95% of the tiime;
2. Are the cut edges straight and are they parallel? The earliest coils can have slightly non-parallel edges, but not much. After 1910 the edges must be perfectly parallel;
3. Were the cut edges cut with a scissors, knife, paper cutter or ?? The cut edges must have the correct appearance from being cut by the coil slicer blades. To observe the correct looking cut edges, use a relatively cheap coil pair such as Scott #390, which is virtually never faked. Or use a cheap pair or larger of a government imperf coil, such as #408H, 408V, 409H, 409V. Get a paste-up pair, which is virtually never faked. Use these for your models for study. Study those genuine cut edges under good magnification (I use 15X, 10X is OK, 15X is better, 20X is OK if you can hold a steady hand, any higher magnification will not be possible to hold steady enough to study); Remember - Even if cut edges look genuine and parallel, the subject coil can easily be made from a genuine imperf coil just be adding perforations, so the next important checkpoint is;
4. Study the perforations. Genuine coil perf holes should have slightly uneven edges, never perfectly round. Perfectly round holes, clean cut, are fake. Again, study the holes of a known genuine cheap pair. For flat plate coils, use #390, 410 or 441. For rotary coils use #448 or 486 or 490. None of these are generally ever faked;
5. Learn what Scott numbers fakes of good coils are made from. This information is found in the Armstrong book or the Schmid book ("The Expert's Book"). Armstrong is better because the info is all basically on one page;
6. Buy the Armstrong book and study the text and pictures. The Schmid book is also good, but Armstrong covers more ground and covers the earliest (pre-1909) coils. Schmid does not.