Having contemplated this issue in depth and given that a lot of information is buried under the sheer size and volume of the PSA Forum thread, I have copied a 2017 post I made there regarding the source of the error in the printing process.
ITS HONESTLY VERY COMPLICATED! And not in any way a sort of a Guttenberg press sort of affair, especially at the volume that Topps was producing what is essentially cardboard packaging. If you can take the time to read all of this and the related links, its very interesting in that you will see the issues that create every type of printing error that we as collectors find on baseball cards.
If you aren't interested in reading, essentially the Thomas and related errors come from either the chemical developing of the photographic negative of the cards for the black plate or the subsequent step of chemically etching those negatives onto the physical metal plate.
The physical metal plates accept ink from the blankets. Due to the issues with the plate itself, the plate did not accept black ink from the blankets (in the areas we find the errors). The plate itself literally is worn out by the production process and must eventually be replaced. Topps would have worn out many plates during any given years production. This wearing out process explains the slight variances between the cards.
IMO - Dave
************************************************** *********
Well first let me say that the four color printing process, a variation of gravure printing, that produces baseball cards, most commercial cardboard packaging and newspaper, is a process that an immense amount of variables play in to. A quick web search will reveal a great deal of scholarly (include complex physics) energy is devoted to perfecting the process. Think fluid dynamics, capillary actions, heat, pressure, static electricity, viscosity, paper quality, ideal chemical properties, speed of rollers, thickness or rollers, tension between rollers, etc.
It defies a short explanation so if anyone is so inclined, check out
printwiki.org/Gravure
Pretty interesting
Another good site which shows common defects in the 4 color process (and basic catalog of every PD in the history of Topps) see
http://www.offsetprintingtechnology....roubleshooting
Gravure printing involves at its heart the engraving of an image onto a printing plate. in the mid-19th century processes were developed to allow for the chemical etching of photographic images into metal. This allowed for the first time for photographic images to be accurately reproduced through press printing (as it was not possible to perfectly reproduce by hand on to the plate). Interestingly today it is possible to physically etch an image with digital technology which can eliminate the chemical process but not possible in 1990.
So at some point Topps photographed and developed the the "F" plate black image. This is when the most likely error occurred that produced the blackless area on the F sheet.
Undeveloped patches are a known problem in developing photographic prints, which would be loosely defined as an area of the negative that has been unaffected by processing solutions. This may have gone unnoticed and the image transferred to the plate.
Also a possibility is that a chemical process is used to etch the photographic negative into the plate. So a perfectly good image can be distorted if the chemical etching process does not go well. I think this is probably what we are looking at with the "F" plate. In this example of the process you can easily see how bubbling or streaking of the sort present on the "F" sheet are easily made during the application of the image to the roller plate.
https://www.youtube.com/watch?v=uwPVKJmElbU
The image in the video is known as a "resist". Problems that occur in rotogravure printing can include the resist not being properly applied. The video above shows a simplified process (usually the resist stays adhered longer to the plate for the image to be properly transferred). The resist usually requires a more involved process for stripping it away from the plate. Sometimes the resist (or parts there of) remain adhered to the plate. This is a very plausible explanation for the errors. To me it seems clear from the type of blackless areas produced and the quantity in which they were produced point clearly to an issue created by the production of the photographic image that produced the resist or its application or removal from the black "F" plate.