Fixed lens focusing

[tigerwillow1]

I compared a couple of new 6mm cameras and found that the image on one was sharper than on the other. I assumed that the focus was set at the factory by robots and would be spot-on. Is it possible that's not the case and readjusting the focus is a sane thing to do? Is it difficult to do?

 

[nayr]

they usually glued in place, it can be done but it'll require you to pick it all off w/a razor knife and then kinda break it loose with some pliers..

you'll also want to print out a test image then you'll have to focus it manually while its half disassembled..

its not for the faint of heart.

 

[Fastb]

Tigerwillow,

I've had luck using several targets for focusing a lens.
- leaves on a tree at medium distance. When not in focus, the leaves get blurry pretty quickly.
- venetian blinds on the windows of our lab had contrast (slats partially open). A crisp edge of the slats was my target. Or the holes that contain the string for raising/lowering the blinds.
- An optical target, made for focusing (google it), pinned on a wall at a medium distance. Then use the zoom feature in Live View . Zooming in seemed to magnify any blurriness due to out-of-focus.

What's also important is to not have "slop" in the lens-to-threaded-housing interface. If it's snug, then twisting for the focus operation requires a little effort to overcome friction. This ensures the lens stays in place after you twist the lens to focus. You do not want the lens to move at all when you go to lock down the lens.

I use a few dabs of silicone. It's strong enough. You don't need to use epoxy to replace the epoxy you picked off to do the focusing. Your cam isn't going on a long, vibration filled trip to go to China on a ship or plane....

Don't forget cleanliness. Micro fiber cloth to clean lens and cam window.

Fastb

Check this thread: Dahua 4MP 25mm LPR Modification

 

[tigerwillow1]

I did the refocus, and it is a pain to do, not only with getting rid of the glue on the threads, but with juggling the case shell, case front, and camera while trying to turn the lens and look at the monitor at the same time. There was enough friction turning the lens that I skipped re-gluing it. I might have made a small improvement, but I'm not positive. I did learn something new, however.

I suspected a focus problem after setting up two 4 MP cameras side-by-side, with one having an obviously sharper picture. Comparing them again after the refocus, I realized that one camera has a sharper image on the right half of the picture, and the other is sharper on the left half. Maybe a side-effect of consumer-level optics?

 

[Fastb]

Yes, good hunch.
The glass lens may not be ground properly.
The optical, curved surface may not be perpendicular to the periphery, which determines how the lens is seated in the lens body.
The threads in the lens holder may mot be perpendicular mounting surface of the lens holder.
The optical sensor may not be perfectly parallel to the pcb, due to movement during the solder process.

When focusing, we used the center of the image. So the L&R would be where the compromise occurs. Or top & bottom.
Our lens holder used a setscrew. We could cant the lens slightly (l to r, up or down) to "fine tune" lens position. Twisting the lens moves the lens in & out. Pushing sideways on the lens angled the lens in a different ddirection/ plane

Fastb

 

[Shockwave199]

I have manually focused lenses too and when possible, it's best to have a second person at the monitor confirming focus and when they say it's good, go in and check it yourself. I couldn't get best focus out at the camera for shit, myself. Ymmv.