DPReview.com is closing April 10th - Find out more

Looking for ideas to build a continuous AF performance testing rig..

Started Jan 15, 2019 | Discussions
(unknown member) Forum Member • Posts: 81
Looking for ideas to build a continuous AF performance testing rig..

I'm looking for ideas, and part suggestions to build an inexpensive rig to test the continuous AF performance of various cameras/lenses. I fully understand it will not perfectly match real-world usage, nor cover every case, but I do wish to make it at-least somewhat repeatable.

At the moment, I'm thinking of a single test that involves a focus-target moving straight towards the camera. A few ideas I was thinking about ...

1) Attach the focus-target on some sort platform that is mounted on a linear rail and can move freely..

2) Have a servo motor to mechanically move the platform at varying speeds/accelerations with ways to start/stop and control the movement precisely. Ideally I'd like to be able to move the platform at around 20-30 meters/second at peak. I was looking at a few DIY CNC rigs that have similar components, but I don't think they need this kind of speed.

3) A way to swap different focus-targets of varying contrast profiles.

4) A way to light the target at varying light levels.

Does anyone have any other ideas on this, or suggestions for specific parts that I could obtain and use for the build?

As a first approximation, I'm going to try to re-purpose a $50 toy train set..

ProfHankD
ProfHankD Veteran Member • Posts: 9,147
Re: Looking for ideas to build a continuous AF performance testing rig..

Kapil Kapre wrote:

2) Have a servo motor to mechanically move the platform at varying speeds/accelerations with ways to start/stop and control the movement precisely. Ideally I'd like to be able to move the platform at around 20-30 meters/second at peak. I was looking at a few DIY CNC rigs that have similar components, but I don't think they need this kind of speed.

Normally, they're doing something like 200mm/s, which is 1/100 what you want. However, 100X is easy to do with a larger diameter connection to the drive belt. Catch is, it will not deal with much weight that way.

Also keep in mind, acceleration is the really big problem, not speed.

There are much beefier steppers and servos if you need to move significant mass.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot SX530 Olympus TG-860 Sony a7R II Canon EOS 5D Mark IV Sony a6500 +32 more
OP (unknown member) Forum Member • Posts: 81
Re: Looking for ideas to build a continuous AF performance testing rig..

ProfHankD wrote:

Kapil Kapre wrote:

2) Have a servo motor to mechanically move the platform at varying speeds/accelerations with ways to start/stop and control the movement precisely. Ideally I'd like to be able to move the platform at around 20-30 meters/second at peak. I was looking at a few DIY CNC rigs that have similar components, but I don't think they need this kind of speed.

Normally, they're doing something like 200mm/s, which is 1/100 what you want. However, 100X is easy to do with a larger diameter connection to the drive belt. Catch is, it will not deal with much weight that way.

Also keep in mind, acceleration is the really big problem, not speed.

There are much beefier steppers and servos if you need to move significant mass.

Yeah, I found a few pre-made rigs that could work, like this one - https://www.ebay.com/itm/163058384667

But the motor doesn't seem to be precise enough from what I could tell. Also, if I just look at the sensor focal plane, and see what 30m/s of subject movement corresponds to, I either need a precise motor that can move fast over a smaller distance (closer focus) , say 2-5m, or a standard motor that can move fast over a larger distance (far focus).

I have dealt with servos at work before, but I don't really know how to go about choosing one, relative merits of belt/screw/gear/rack and pinion, etc so any help is greatly appreciated!

ProfHankD
ProfHankD Veteran Member • Posts: 9,147
Re: Looking for ideas to build a continuous AF performance testing rig..
1

Kapil Kapre wrote:

ProfHankD wrote:

Kapil Kapre wrote:

2) Have a servo motor to mechanically move the platform at varying speeds/accelerations with ways to start/stop and control the movement precisely. Ideally I'd like to be able to move the platform at around 20-30 meters/second at peak. I was looking at a few DIY CNC rigs that have similar components, but I don't think they need this kind of speed.

Normally, they're doing something like 200mm/s, which is 1/100 what you want. However, 100X is easy to do with a larger diameter connection to the drive belt. Catch is, it will not deal with much weight that way.

Also keep in mind, acceleration is the really big problem, not speed.

There are much beefier steppers and servos if you need to move significant mass.

Yeah, I found a few pre-made rigs that could work, like this one - https://www.ebay.com/itm/163058384667

But the motor doesn't seem to be precise enough from what I could tell. Also, if I just look at the sensor focal plane, and see what 30m/s of subject movement corresponds to, I either need a precise motor that can move fast over a smaller distance (closer focus) , say 2-5m, or a standard motor that can move fast over a larger distance (far focus).

I have dealt with servos at work before, but I don't really know how to go about choosing one, relative merits of belt/screw/gear/rack and pinion, etc so any help is greatly appreciated!

I'm not going to claim expertise here, although I've certainly used/built plenty of cheap motion control systems. I think it all comes down to:

  • Desired object travel
  • Desired object sustained speed (and repeatability constraints)
  • Desired force -- f=ma, i.e., what moving mass (including friction) and acceleration

To put it bluntly, if the object is a target that weighs less than 50g and is hanging from a cord looped around a pulley, you could probably get away with a really wimpy motor. If you're moving a 5-pound target on a greased rail, you probably need an industrial servo or somesuch.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot SX530 Olympus TG-860 Sony a7R II Canon EOS 5D Mark IV Sony a6500 +32 more
D Cox Forum Pro • Posts: 32,980
Re: Looking for ideas to build a continuous AF performance testing rig..

I suggest abandoning the straight line and starting with a record-player turntable, preferably one with a full 12 inch diameter.

A suitable target might be a cylinder of stiff paper of about 3 inch diameter with a suitable black and white pattern on it.

 D Cox's gear list:D Cox's gear list
Sigma fp
Keyboard shortcuts:
FForum MMy threads