TESTING VISUAL TELESCOPIC DEVICES About Wells Research Joined TRIOPTICS mid 2012. Currently 8 employees Product line compliments TRIOPTICS, with little overlap Entry level products, generally less expensive TRIOPTICS sales force is our primary distribution path Defining Visual Telescopic Devices This document addresses instruments with the following characteristics: Are intended for viewing objects at moderate to long distance. (10 yards to infinity) Have one or two eyepieces which a human user looks into. Have outputs in the visible spectrum. Have magnification of typically 1X to 10X, occasionally higher. May involve a reticle for visual reference. This list includes monocular telescopes, spotting scopes, riflescopes, and binoculars. The devices may be simple telescopes, or be night vision devices (NVD) with an image intensifier tube. In the sections that follow, we'll represent these devices with simple schematic sketches. For example, in the sketch below, the blue arrows represent rays of light from a single point on a distant object: Fig 1 Fig 2
Measurements Most customers want to make some or all of the measurements in this list Image quality: MTF Limiting resolution and 3-bar contrast Basic optical parameters: magnification, FOV, boresight error Additional optical parameters: field flatness, distortion Color aberrations, such as axial color and lateral color. Exit pupil size and location Transmission, spectral transmission (or gain if NVD) Unique issues (e.g., the number of dead fibers in an NVD device) Table 1 The basic measurements are easy, but in our experience, real-world devices have complications: Image quality almost always varies across the FOV. The plane of best focus is generally not flat, and often has astigmatism. If the exit pupil is large, then image quality will almost certainly vary depending on location of users eye pupil within the exit pupil. Distortion may even vary depending on location of users eye pupil relative to exit pupil. When there is a reticle in the device, specialized tests are required such as parallax, tip and scale factor. NVD devices require light shielding, and very large dynamic range in illumination source. Special issues arise when there is a zoom feature in the device: Besides magnification change, there may be issues of backlash or image shift. Table 2
Measurement examples Wells Research PixelScope software is a powerful, easy use platform for making the measurements in Table 1. For more information please request the PixelScope spec sheet or contact your local TRIOPTICS representative for a demonstration. 2 1 3 Fig 3 Basic approach to testing Over the years many different approaches have been used to make the measurements in Table 1. However, based on our experience we recommend the following approach: Fig 4 (A) is a collimator, which simulates light from an image far away. (B) is the visual device under test (C) is a video telescope, used to view the virtual image created by instrument B. The green arrows indicate that the collimator and video telescope can pivot for testing at of-axis locations.
In almost all cases we recommend a motorized video telescope whose diopter focus can be adjusted under computer control. This capability permits extremely precise measurements of measurement of field curvature, astigmatism and parallax. Here's a physical test bench that corresponds to Figure 4: Different types of monocular devices Fig 5 For monocular devices like riflescopes or spotting scopes the Riflescope Test Station shown in figure 5 is a good solution. Collimators with clear aperture up to 125 mm are available. Collimators have adjustable focus, and can simulate images at infinity or as close as 25 yards. The video telescope shown above, has a focus range of +/- 5 diopters, and up to +/-8 diopters is available on special order. Dental and surgical magnifiers are a special case. Working distances are much shorter, so collimator is replaced by a small back-lit target. Fig 6 Fig 7
Head Mounted Devices are another special case. Because the devices are not rotationally symmetrical, separate tip and tilt stages may be appropriate. The small white device above is a Private Eye an early head-mounted display which lacked see-through capability. While not shown, it's possible to add a collimator for testing see-through devices as well. Fig 8 Folded devices, or devices with unusually large FOV requirements, can best be handled on an optical bench: Fig 9
Approaches for binocular devices Binocular devices bring unique challenges. Many approaches are possible, but in general we recommend separate test stations for image quality and for collimation. Image quality testing of binocular devices For Inline devices, it's possible to use the type of bench shown in Fgure 4, 5. It's only necessary to add a X direction slide underneath the binocular cradle. The slide allows shifting between the left and right eye path. Both motorized and manual slides are available from Wells Research. For offset devices such as porro-prism binoculars a second slide is often necessary: Fig 10 Fig 11 It is possible to use these fixtures for collimation testing as well. However, to make accurate collimation measurements an unusually precise cross slide is needed. As a result we generally recommend a separate fixture dedicated to collimation testing. The reason an ultra-accurate cross slide is required is not obvious, so it's worth a quick explanation. Let's say we are testing 15X binoculars, and let's say the cross slide has a wobble of 1 arc-minute between the L and R eye positions. Or, if you prefer, assume that the rubber casing on the binoculars has a tiny bit of compliance. When switching from L to R eye, perhaps there's a slight bump as you reach the end position. If so, it's easy to imagine that the binocular body could shift a tiny amount in the cradle (let's say 0.0005 inch)
In either case, the binocular body rotates by 1 arc minute as you move from L to R eye path. That doesn t sound like a lot, but remember that the binocular body has power. With our hypothetical 15X binocular, the rays coming out of the eyepiece will rotate by 15 arc minutes when the body rotates by 1 arc minute. Now that's a lot. As a result we recommend the use of a dedicated dual-path fixture to measure collimation. The results are more robust, the cycle time is faster., and the end result is generally more satisfactory. Collimation testing of binocular devices The simplest approach uses a pair of autocollimators and a large flat mirror. This approach is simple, fast, and exceptionally accurate. It is useful both for devices with straight-through optical paths and also devices with an offset. This setup can also be used for basic image quality analysis. Of course the measurement is only on-axis, and because of symmetry some image errors cancel out 1. Nevertheless, this is still a valuable trip wire test for manufacturing problems. Unfortunately, the simple mirror fixture shown in figure 12 cannot detect image rotation or magnification. If these must be measured, there are two approaches 2. Fig 12 If an off-axis bench of the type shown in fig 4, 5 is available in the plant, then this bench can be used for rotation and magnification measurements. If no other bench is available, or if its necessary to make rotation or magnification measurements on a very large number of units, then it is possible to add accessories to the basic mirror bench. 1 Spherical aberration and other even image aberrations cancel out due to symmetry in this setup. Explanation of this phenomenon is beyond the scope of this document. Nevertheless, (in the author's experience) binoculars don't fail final test because of too much SA. They fail because of problems caused by lens centration and similar manufacturing issues. 2 It's also possible to argue that prisim quality should be monitored at the component level, rather than waiting for final test.
Image rotation can be measured by adding a pair of 90 degree prisms to the basic mirror bench: Fig 13 If magnification must also be measured, Then it's necessary to add two collimators, Fig 14 Fig 15 Summary: Testing visual telescopic devices has unique challenges. If you would like to take advantage of our years of experience in testing these devices, please contact Wells Research or your local TRIOPTICS representatibve.
APPENDIX Some Frequently Asked Questions: Q: Are off-axis pivots really needed in the image quality test fixture? What if I only want to test on-axis? A: This is certainly an option, especially if the equipment is to be used for testing to a specific on-axis specification and a large number of units must be tested. The off axis pivots may be added later if desired. Wells Research equipment is modular, so we have the ability to configure a system to specifically address customer requirements. Q: Can collimation fixture in figure 12 be used for image quality testing? A: Yes, it's possible to detect image problems with the collimation fixture. However, since the test is double-pass, the image on the video monitor is not exactly what you would see looking though the device. Because of this it's not possible to make a valid MTF measurement in double-pass. (This is a subtle topic that warrants its own white-paper!) Nevertheless, it is absolutely possible to use the collimation fixture as an early warning station to detect image problems. Q: Can the image quality bench in figure 10 be used for collimation testing? I only need 1 arc minute accuracy. Surely the cross slide is that accurate. A: Yes, but when testing high power devices, there is a complicating factor. If the telescope under test is 15 power, then the cross slide must be ~15X more accurate than the accuracy required at the eyepiece. Thus the accuracy requirements are not as simple as the 1 arc minute requirement suggests. Q: Why do I need a pivoting collimator? Why can't I use a wide angle collimator that covers the entire field of view? A: This is possible in concept, but in practice it is surprisingly expensive to build a wide angle collimator with sufficiently good wavefront quality over a wide aperture. Also, the pivoting collimator has a sublte advantage: because the collimator pivots, the image quality from the test fixture is exactly the same at the edge of the FOV as it is on-axis. Q: Similar question about the video telescope: Why can't I use a wide angle lens that covers the entire field of view? A: Similar answer: it is surprisingly expensive to build a wide angle lens with good enough wavefront quality across the entire image plane. Also, to cover the full eyepiece field of view a very large number of pixels would be needed. Nevertheless, in a few situations a single wide field camera may be appropirate.