Good news, I managed to find all my old FED data including photos of my collection and the old graphs!
So I should be able to restore everything, might take a while though. Page 1 should be fine again, let me know if anything does not add up, I had to make a few changes.
The camera/lens pair graphs above are updated as well.
Now what strikes me most is that though the number of FED S camera/lens pairs has doubled since the first graph, virtually nothing changed. The model I proposed in 2013 is still perfectly valid.
As I had forgotten everything I did all calculations from zero and only noticed afterwards I got virtually the same results as back in 2013.
I'll show you:
Sorry for this cluttered graph, take your time and I hope you'll understand.
Basically we see a fundamental difference in the way lens serials were assigned to FED S 'c' cameras than to FED S 'd' and 'e' ones.
As I concluded in 2013 the ratio is 1:11 for 'd' models and this is highly certain. No matter how many cameras we add I doubt we'll end up anywhere else than between 1:10 and 1:12. Also the model correctly predicts the number of manufactured cameras and lenses.
The fundamental questions are:
1) How did they manage to achieve this?I looked into the serial distribution of FED-S to find patterns of how cameras were picked from the range of FED1 cameras to become an 'S'. Though the distribution is most likely not completely random I have not come up with any clues. What I can rule out is that regularly larger batches were converted to FED S. We'd see a stair-like distribution which we don't - it seems completely linear on the resolution we have. However from the current data we cannot see what is going on in the range of dozens of cameras. Hundreds at best. And the model suggests roughly every 10th FED1 was a FED S.
What we can also rule out is independent production of cameras and 50mm f/2 lenses and then picking random cameras for FED S conversion. This does not work, and would lead to random blob-like distribution, not our almost but not completely perfect line of 'd' cam/lens pairs.
I normally work with environmental data and I have never seen a distribution like this. It certainly shows well-planned human work but also not some perfectly executed industrial process.
If I would not color them in the graphs you could not discern 'd' from 'e' at all.
Look at the 4 lenses perfectly on the regression line and a fifth only a bit above it.
How can we explain this if the factory was evacuated to Siberia? Impossible to retain exactly the same production ratio under such circumstances. Fact is - the latest known FED S'e' pair #181306 / 33408 (Fed plate mounted Sonnar), the rightmost blue dot, fits the regression line perfectly. I find it hard to believe that anything fundamentally would have changed in production up to this serial range. The 'outliers' are so far out and random seems strange they would intermittedly use these old lenses in the factory and then resumed their perfect cam:lens ratio shortly after. I do not think these lenses are original.
To me it seems like they planned beforehand which lens serial was to be assigned to a certain camera serial range (in the hundreds), but not to an individual camera. Meaning 50mm f/2 lens production followed FED1 production, not FED-S production.
Also we do not see lens batches at all. It appears lens output was strictly coupled to FED1 production. If camera serials rose by ten, lens serials rose by one. And this for the entire 'd' range and seemingly well into 'e' production without any disturbances. Another amazing detail.
In contrast something fundamentally different was going on during 'c' production. the pairing of cameras and lenses was far more fuzzy and roughly followed an 1:7,5 ratio. But maybe even less down to 1:6? It seems the beginning and end of FED S'c' productuion was chaotic. In order to arrive at around #54.000 for lens 20.000 initially more cameras had to be picked to become 'S' than later on. Either they alerady had lots of 50mm f/2 lenses when FED'c' production started or they could not keep up FED S procuction and had to settle for fewer ones? Also towards the end there appear to have been serious problems. We have few FED S cameras in the range 80.000 - 95.000 and they have odd lenses. Likewise the expected lenses in the range 24.000 - 25.750 we know of ended up with much later 'd' and even an 'e' camera.
Now regarding the curious number of 33.5xx 50mm f/2 lenses:
I always found it weird to plan the maximum number of components beforehand. Keep in mind that in contrast to all other items the 50mm f/2 was not an 'accessoir' to be bought separately but was sold with the FED S. So we can assume that if more cameras than planned were produced they also had to exceed the lens number.
However lens 30.000 falls in the mid 'd' range and we see nothing unusual there. Anyway I think there are way stranger things about the FED S than the 'excess' 3.500 lenses.
