Having a 1966, as well as my '67, this thread peaked my interest. However, before getting into details, we need to get a little nomenclature straightened out. What you have noted as a "Part Number", I am assuming (?) that you are getting that from the numbers cast into the back(inside) of the Headlight Door. These numbers are not Part Numbers. To explain further, there are basically three different numbers encountered with parts. First is the number given the part when it is designed. This is also called the Engineering Number (also the Factory Part Number or Assembly Line Part Number). Second is the Casting Number. This is the number (not a part number) either cast into, stamped into, printed on, or printed on a sticker applied to, the part. Many times, although not always, these are the same as the Engineering Part Number. And finally, is the Service Part Number. This is the number that is assigned to the part to be stocked in the Dealers' Parts Departments for use in servicing parts after the car left the Factory. The Service Part Number is what is most commonly meant when people refer to "Part Numbers", unless "Engineering" is added, or unless "casting number" is specified (note here that only "casting number" is correct, not "casting part number".
That all being said, back to the issue at point. Being curious about which casting numbers my 66 had, I knew that I still had my originals (which show evidence of the car's original Ivy Green color in the form of overspray inside of the Headlight Door. When I purchased my 66 from its second owner in 1976, it exhibited no evidence of having been repainted. I looked at the casting numbers and they were C5ZB-13064-C (LH) and C5ZB-13052-C (RH). I then went to my 1966 Electrical Assembly Manual (Headlight Doors are listed in the Electrical Assembly Manual rather than in the Body Assembly Manual as one might think), to see what the Engineering Number was shown as ; in this case, they were the same as the casting numbers, C5ZB-13064-C and C5ZB-13052-C. The appropriate drawing in the Electrical Assembly Manual was dated "3/30/66, Supersedes 12/11/64".
I then checked the Part Numbers (Service Part Numbers) shown in my 1966 Parts & Accessories Catalog (single year MPC, if you will) ; C5ZZ-13064-B (RH) and C5ZZ-13064-A (LH).
Then, I remembered that some years ago (in the late 80's I believe), I had picked up a spare set of Headlight Doors. I was of course curious to see if they were the same. Interestingly, they are the C5ZB-13064-A/ C5ZB-13052-A casting number version. At some time after picking these up (I can't remember from a car or at a swap meet), I had bead blasted them and primed them, and put them away for the next "go around" on my 66, so they are easy to tell the difference from my originals.
The first thing I did was to check to see if both exhibited the slight "peak" to match the fender contour "peak" ; they did. Then I traced each onto the same 11 x 17 blank paper in different color inks (a little difficult since they are not flat). As best as I could tell, they were identical except the "-C's" appear to have a slightly longer length (1/32" to 3/32") toward the radiator on each side. I would think that the 1/32" side could be within "casting tolerance", but the 3/32" side would seem too much for that. Although, since I had to "rock" the casting to complete the tracing, it could just be my measurements.
Next, I looked at the back side of the casting on each. The "-A's" seem to be "beefier", with additional stiffeners around the perimeter, and slightly larger bosses for the attaching screws. However, in weighing each, the "-C's" were actually heavier by 3 ounces (1 lb - 6.4 oz vs 1 lb - 9.4 oz). I have attached pictures of both castings (remember, the red oxide primered ones are the "-A's").
My next question was "when did this change occur ?". Typically, like Engineering Numbers and Service Part Numbers, casting numbers' suffixes are alphabetical ; B after A, C after B, etc. Since I have neither the 1964-1/2 or the 1965 Electrical Assembly Manuals, I couldn't check either for the Engineering Numbers. So, I went to check on the Part Numbers. I checked my 1965 and 1966 single year MPC's (each printed in January of their respective years). Both showed the same Part Numbers of C5ZZ-13064-A & -B.
Next was to check my OSI (Obsolete-Superseded-Interchange) catalogs. Unfortunately, after January 1964 (before Mustangs were introduced), my next one is not until July of 1965. However, in the July 1965 OSI, it shows that C4ZZ-13064-A & -B were to be "mw" (mixed with) C5ZZ-13064-A & -B. Thus, we can only say that sometime between January 1964 and July 1965, the C4ZZ's were to be substituted with the C5ZZ's. Just a little more clarification before "moving on". Generally speaking, when a part is noted in the OSI catalogs as "mw", means that the part will function and fit the same as the part it is being "mixed with", but there may be some minor differences, particularly in appearance. If a part is shown as "r/b" (Replaced By), this usually means that the part is virtually identical to the part it is replacing, with possible "internal differences". In the specific case of the Headlight Doors, I really don't know why Ford chose to designate "mw" as opposed to "r/b", as, at least to me, they both appear to be identical from outward appearance.
I realize that the above doesn't really answer the questions as to why both fmr187 and sglbbs have/had seemingly opposite problems (possibly due to when the change occurred), and sglbbs car being an earlier build date and the "A's" still being used, and fmr187's being later requiring the "-C's" ?), but maybe it will lead to more information as to "why".
Finally, in order to make this comparison complete, it would be beneficial to this "story" if someone with the 1964-1/2 and 1965 Electrical Assembly Manuals would Post the Engineering Part Numbers shown in each for the Headlight Doors, along with the Date shown on those drawings.
I hope that I haven't confused the issue, but wanted to give my observations.
Bob