Building a Home-Made Mechanical-Focus Camera

[armstrong]

Hi folks

I have an amateur interest in photography, but a significant hobby interest in optics and electronics.

I have decided I want to build working elements of an auto-focus camera. This will obviously require several stages. What I would first like to get done, is the focusing element of auto-focus cameras.

Is there a webpage or a book detailing the inner-workings of an auto-focus camera? what kind of light is used (infra-red?). what exactly happens - does the camera emit light, then has a sensor that detects its reflection, and from then, judge distance?

i'm thinking a very simple model auto-focus (none of the fancy stuff with multiple focus etc.). In fact, if there's a more basic point-and-shoot model i could first read/learn about, then that will be fine too (e.g. what are the optics of a simple point-n-shoot?

Thanks again. Please move this post to the appropriate forum if this isn't the right one.

 

[Alpha]

I don't know much about engineering. I know a little about optics, though not the particulars of AF lenses so much. Off the top of my head, I was thinking that it might be easier to use a servo to move the film/sensor plane than to internally focus within a lens. If you're interested at all in the engineering of that, you could try looking up patents for the Rollei 6008AF or Contax AX cameras.

 

[armstrong]

thanks for your reply. i don't know much about the history of camera development, but i'll definitely look up those cameras.

can anyone else chime in on this question? i'm interested in the particulars of the auto-focus element of auto-focus cameras. what type of light is emited? what is the sensor? what type of chip determines the blur?

 

[Alpha]

AFAIK, the focusing system has little to do with the aperture or blur as you call it, except that it focuses wide open and then stops down, when you're shooting below maximum aperture. If you're interested in that, you'd do well to read up on electronically controlled aperture.

Without being able to go into too much detail about the engineering of such optics, I can tell you that AF works by moving an element or elements inside the lens body, such that all of the light converges at a point where the film/sensor plane is. The math behind it is relatively simple. If the light converges in front or or behind the film/sensor plane, then the image will be out of focus. The same rules apply to nearsightedness and farsightedness in human eyes.

 

[usayit]

AF (at least the early systems) had sensors that would adjust the lens to establish maximum contrast.

In the past there were two types of AF systems; ones that relied on data collected through the lens and others that would determine the range of an object and set the lens accordingly.

One example of the later is the ol'sonic polaroid systems. They sound waves to determine the distance of an object and focused the lens accordingly. They essentially automated how a rangefinder operates.

I personally would guess that it is easier to impliment this type of AF.

 

[Helen B]

There are also split-image systems that work behind the lens, as well as the contrast systems already mentioned.

Rangefinder-type systems are still in use. Any camera that has a leaf shutter that is normally closed (ie most compact film cameras) must use such a system. High-end motion picture cameras may also use a rangefinder connected to a lens focusing servo.

There was an old Bell & Howell movie camera that used simple triangulation: you pointed the camera at the ground, and the lens was set by a pendulum. This assumed that the camera was at a certain height, and that the ground was level.

Best,
Helen