Building a frankenstein camera | Part 3: The Flange Distance

The Flange Distance

With the rangefinder question more or less settled, it was finally time to actually open the FED 2 up and see what I was working with. I haven't stripped the internals out yet - for now I just opened the back and took a caliper to it. That alone was enough to get a number I didn't like: from where the new digital sensor would need to sit, there's only about 5mm of depth left to the back of the housing. That might just barely be enough for the Sony sensor stack and its mainboard, but there's no way it also fits the capture card, the Raspberry Pi Zero, and one (or two) batteries on top. I want to use every cubic millimeter inside this shell as efficiently as possible, so 5mm wasn't going to cut it.

What Even Is Flange Distance

Before figuring out how to fix that, I had to actually understand what I was measuring. Flange distance (also called flange focal distance, or register) is simply the distance from the mount itself - on a screw mount like ours, right where the first thread starts - straight back to the sensor or film plane, measured along the optical axis. Every lens is designed assuming a specific flange distance for its mount - it's what lets the lens focus correctly all the way out to infinity. Get that number wrong, even by a millimeter, and infinity focus is the first thing you lose.

Small side note: the mark you sometimes see etched on top of newer camera bodies - a circle with a line through it - is the sensor/film plane indicator. That's the mark you'd actually measure flange distance from if you don't want to open the body up.

For the FED 2, which uses the M39 Leica Thread Mount (LTM), that number is fixed: 28.8mm. Every LTM lens ever made assumes there's exactly 28.8mm of air between the mount and the film (or sensor). For context, here's how that compares to other common mounts:

Calculating Adapter Length

The rule for a simple, dumb (optics-free) adapter is:

Adapter length = Lens's native flange distance − Body's flange distance

For that number to make physical sense, it has to be zero or positive - you can always add material between a lens and a body to push the lens further out, but you can't remove material to bring it closer than the body's own mount already sits. So an adapter only works, without corrective optics, when the lens's native flange distance is longer than the body's.

That's exactly why Sony E-mount bodies are such a popular target for lens adapters: at 18mm, its flange distance is shorter than nearly every lens mount ever made - LTM, Leica M, M42, EF, F, K, you name it. Plug any of those into the formula and you get a positive number every time, meaning a simple machined spacer ring is all it takes to adapt almost any lens in existence.

Not Enough Room, So We Move the Mount Instead of the Shell

Here's the constraint I kept running into: I don't want to drill new holes or replace the housing. That rules out simply extending the shell backwards to make room for the electronics stack. But there's another way to think about this problem. If I move the sensor forward - closer to the bayonet, physically shallower inside the shell - I free up roughly 2 to 2.5cm of depth at the back. That's a huge amount of extra room for the mainboard and battery.

The catch is obvious once you think about the 28.8mm rule: if the sensor moves forward but the mount stays exactly where it is, the flange distance shrinks and every lens loses infinity focus. So the mount has to move forward too, by the same amount, extending out past the front of the shell rather than sitting flush with it. Rather than extending the case to the back, we extend it to the front. Net effect: the sensor sits much shallower inside the body, but the distance from the new, extended mount face to the sensor is still 28.8mm, so lenses still behave exactly as they should.

Why I Can't Just Buy an Adapter for This

My first instinct was to look for an off-the-shelf spacer ring to do this. Lensless (dumb) adapters exist for all kinds of mount combinations, and going by the formula above, that's exactly what one is - a ring whose thickness equals the gap between two flange distances. But every adapter on the market is built for a known pair of standards - LTM-to-Sony-E, M42-to-EF, whatever. If I would buy an adapter where the thickness is slightly off from my calculated distance - I am going to lose infinity focus, because the lens is now focusing to a plane that doesn't line up with the sensor anymore.

My situation isn't "adapt standard A to standard B" - it's a bespoke, arbitrary extension to reclaim internal depth, with a distance that no manufacturer has ever needed to standardize. Nobody sells a ring for that, because it isn't a real lens-mount conversion, it's a hack specific to my shell. Which meant designing and building my own.

Worth noting: this "losing infinity focus" tradeoff is exactly the trick behind extension tubes, and it's done on purpose there. Add a dumb spacer ring of arbitrary thickness between a lens and a body, on the same mount, and you push the lens's whole focus range closer - a lens that normally focuses down to, say, 60cm might now only focus between 5 and 15cm. You've sacrificed infinity to gain macro range. It's the same physics as my problem, just aimed in the opposite direction on purpose, rather than as an unwanted side effect.

The M39 Naming Trap

While researching this, I ran into something that would have bitten me hard if I hadn't caught it: "M39" doesn't reliably mean the same thread. The FED 2 (and Leica, and Zorki) use LTM - the Leica Thread Mount - which is a Whitworth-derived fine thread. But Zenit also called their mount "M39," and it is a completely different, incompatible thread standard underneath the same name.

Same nominal 39mm diameter, same marketing name, completely different geometry. Thread a Zenit M39 lens confidently onto an LTM body (or vice versa) and best case it's stiff and wrong, worst case you cross-thread and damage the mount. Definitely something to know before ordering thread taps or trusting a random "M39 adapter" listing.

A Quick Detour: Why Whitworth on a German Camera?

This one bugged me for a bit. Whitworth is a British standard - why would Leitz, a German optics company, build their lens mount around it? As far as I can piece together, it comes down to timing rather than nationality: in the early 20th century, before ISO metric threads were the international default, Whitworth-family fine threads were already a well-established precision standard used widely across the optical and instrument-making industry, Germany included. Leitz wasn't picking "the British thread" so much as picking an existing, trusted fine-thread standard available to precision workshops of the era. ISO metric threading only became the dominant standard well after LTM was already locked in, so the mount just never got redesigned.

Confirming the Exact Thread I Need

With the naming trap out of the way, I needed the precise numbers, not just "Whitworth-ish." Digging through references and cross-checking against the Fusion 360 thread I ended up building, the LTM/M39 spec comes out to:

• Nominal diameter: 39mm
• Pitch: 0.977mm (26 TPI)
• Flank angle: 55° (Whitworth profile)
• External major diameter: ~38.85mm
• Internal major diameter: ~39.1mm

Those numbers are what actually went into the custom thread I modeled. Read my post about adding custom threads in fusion 360 here: Fusion 360 - Create Custom Threads.

Designing the Extension Tube

Modeling-wise, I first reached for Plasticity, since it sits in that nice middle ground between a organic 3D tool like Blender (which if often use as technical director) and a full parametric CAD package, and it's what I've been using for the housing tests i wrote about in my last blog post [[Building a frankenstein camera - Blog Post 2 - The Rangefinder Problem]]. But building an actual, correct thread in it turned out to not be straightforward - no proper thread feature, so I would have had to hand-model the whole helix. Given I now know exactly which thread standard I need and want it to be reusable, I switched over to Fusion 360, where I could define LTM/M39 once as a real, selectable thread standard and just apply it going forward.

The part itself is simple in concept: a ring with the LTM thread cut internally on one side (screws onto the FED 2's original mount) and externally on the other (so lenses screw onto the new, extended face), with a wall of exactly the length needed to reclaim that 2 to 2.5cm of depth. Simple geometry, but it only works if the thread itself is dead accurate - which is the entire reason I went down the custom-thread-standard rabbit hole in the first place.

Extension tube M39 to M39

Adapter M39 to M42

Print First, Machine Later

For the first version I'm 3D printing the tube. One thing worth flagging for anyone doing the same: don't forget to use classes/tolerance in Fusion 360's thread definition when your part is going to be printed. A class adds a bit of extra clearance between the thread spirals, which matters a lot for FDM printing - printed threads are never as dimensionally accurate as machined ones, and without that clearance the ring simply won't thread onto the body or accept a lens. Once I've confirmed the fit and the flange distance actually lands where it should, I'll look at getting a proper aluminum version CNC machined - printed plastic is fine for proving the geometry, but I'd rather trust metal against a lens mount long-term.

Next up: getting this ring printed, threading it onto the FED 2, and finding out whether that reclaimed 2 to 2.5cm of depth is actually enough to fit the new Camera internals and a battery - or whether I'm about to learn a whole new lesson about power delivery in a tiny space.