Lens centering: Should the mount move down or up in this case?

[Fedor Karpenko]

Hey everyone! Want to double check something with you guys:

I got a new front for my K3 that has a PL mount on it. The centering is a little off so my s16 lenses end up having some vignetting. (If you're looking at the ultra 16 image area then the vignetting is pretty bad, but in terms of the s16 image area it's not that horrible)

I'm able to get a new front for the camera with adjusted mount placement. From my understanding, the mount would actually need to move upwards a bit, not downwards, right? I created this flipped version of the image to see the way the image lands on the film out of the rear of the lens, along with an approximate image circle of the lens based on the vignette:

I estimated that the mount would need to move upwards by about 1mm to get it centered up on the Y axis. If you're just looking at the s16 area, it seems like the X-axis centering is actually quite solid. (not perfect, but probably not worth tweaking since I'm only going to use the s16 area)

Would really appreciate some input!

[Dom Jaeger]

Seems about right.

[Robert Hart]

Before you set about keyholing the attachment screwholes or other surgeries to move the optical centre of that lens mount, it might be worth checking if there are any shims that have folded or creased at the screwholes and tilted the lens off-axis. 

A practical test for off-axis tilt might be to open the lens iris up to the maximum and look for soft focus on an upper or lower frame edge. You would need to have a flat textured surface or even better, a proper chart as a focus target with the camera set up dead level and square-on to the focus target. 

Edited December 11, 2022 by Robert Hart
error

[Fedor Karpenko]

4 hours ago, Dom Jaeger said:

Seems about right.

Thank you!

[Fedor Karpenko]

1 hour ago, Robert Hart said:

Before you set about keyholing the attachment screwholes or other surgeries to move the optical centre of that lens mount, it might be worth checking if there are any shims that have folded or creased at the screwholes and tilted the lens off-axis. 

A practical test for off-axis tilt might be to open the lens iris up to the maximum and look for soft focus on an upper or lower frame edge. You would need to have a flat textured surface or even better, a proper chart as a focus target with the camera set up dead level and square-on to the focus target. 

I'll double check, although I'm pretty sure there's no tilt. I checked the FFD with a depth gauge and it was pretty spot on across the entire gate, I'm guessing if there was a tile then the depth gauge would give me different readings in different edges of the frame, right?

[Robert Hart]

You may not detect a tilt with flange to focal plane manual measurement. The distance added or subtracted by a tilted mount or lens tail to produce an off-centred image on a 16mm sized image area would be very small. Some swarf or grit between the mount and the camera body might be all that is needed to cause it. How to find it is a challenge. A test roll with the camera set up with the mount face perfectly at 90 degrees to a focus chart with the chart lit for the lens to be set at wide-open aperture or projection of a texture on a groundglass panel in the gate through the lens may show softness in edges or corners indicating an off-axis tilt. 

That assumes the centering of the lens mount was done correctly in the first place. With the small 16mm image frame there is not much room for error. What you have there is about a 0.5mm error. The former Soviet K3 may not have been an exemplar of precision and alterations since may not have been either. Another consideration is whether the claw crank mechanism has become worn and the claw has begun to strike short. For a 0.5mm error I would expect a LOT of mechanical noise and clicking of misregistering when the claw engages into or disengages from the sprocket holes in the film.

If the option of sending the camera to a tech is not viable, my personal preference would be to selectively loosen but not remove the fastening screws across the lower semi-circle of the mount-to-body attachment. There may be enough compliance in the body for the lens to tilt upwards and carry the optical centre axis upwards slightly. To check this without rolling film through the camera would require making a groundglass focus target to fit into the gate and reverse projection with the camera very firmly mounted to a solid base.

Somebody else can correct me here. - An off-axis movement of the lens, either a tilt or simple lateral movement carries the image circle across the focal plane in the same direction as the movement of the lens. 

If your test restores the image centre, then my personal preference purely as a backyard engineer would be to remove the mount from the body, check the matching faces for swarf and grit, remove the contamination, reassemble and test again. If that does not work, then my next personal preference would be to find the thinnest possible scrap of shim material about 5mm in area and fit it between the mount and body at the lowest point on the lower half-circle of the mount, then refasten all screws.

This will tilt the lens axis upwards and raise the centre of the image circle at the focal plane. If this succeeds and you are content with having an uneven loading of the structure then no more needs to be done. It is more likely you will find that the error has become over-corrected. Edge or corner softness may have been introduced. 

My personal preference if the image remains sharp across its entire area would be to polish the upper half circle of the mount to body contact face on the body with a tightly bunched bud of steel wool, clean, re-assemble and retest to see if there is any difference. 

Take care to plug the centre hole against loose fragments of steel wool entering and causing trouble. 

Please take heed of better and smarter people than I if they respond here. There are correct ways of doing this, not my kitchen-table methods.

Edited December 14, 2022 by Robert Hart
error