Building an 1850s-style Direct Ophthalmoscope

For over 150 years, people have engineered ways to view the back of the eye, which is called the fundus. Most of us are familiar with these fundus viewing tools from routine eye examines. The simplest device is called a direct ophthalmoscope. They're simple optical devices, but they can provide an amazing view of the fundus. I decided to build one based on the original design from 1851 that uses candlelight for illumination.

50:50 beamsplitter:

Dowel rod

Candle

3D printed parts

Correction lens (optional)

The human eye is an optical system. The cornea and the crystalline lens focus light on the retina, with most of the focusing power coming from the cornea. The iris controls the amount of light entering the eye, acting as the aperture stop.

I've added a diagram showing the position of the entrance pupil relative to the iris, as well as the marginal ray, chief ray, and exit pupil. This figure is from one of my favorite textbooks: Michael Kaschke, Karl-Heinz Donnerhacke, and Michael Stefan Rill. Optical Devices in Ophthalmology and Optometry: Technology, Design Principles, and Clinical Applications. 2014. https://onlinelibrary.wiley.com/doi/book/10.1002/9783527648962

The optic disk and retinal blood vessels that enter and exit through it are clearly visible in the fundus photo.

Now that we have a bit of anatomy to give us context, we can get to the challenge of viewing the fundus. If not enough light is reaching the fundus to see it, why can’t we just shine a flashlight into someone’s eye?

The flashlight is too bulky for the illumination and detection pathways to overlap, because the iris is so small, and it’s made worse as the iris constricts. By overlapping the pathways or simply moving them close enough together, it’s possible for them to both pass through the pupil and overlap on the fundus.

A common design today has an illumination source focused to a point, which is then reflected off a mirror towards a patient. A viewing aperture is placed above the mirror. That’s it. When the physician lines up their pupil, the viewing aperture, and the patient pupil, they can see a red glow coming from the patient’s eye. It’s creepy.

I took apart a low-cost direct ophthalmoscope. The first thing you see is a ring of small plastic lenses that rotates, a physician selects the lens that compensate for the patient’s refractive error. Moving the ring of lenses aside, you can see the light source and a small illumination mask above it. There are three plastic lenses for imaging the illumination mask to the fold mirror. Splitting the power with three lenses helps minimize aberrations. The fold mirror is glued right below the viewing aperture.

Safety is really important in medical imaging, so you’ve got to use samples before real human testing. Shining light into your eyes with a homebuilt device (even with a candle) and no medical expert shouldn’t be taken lightly. There’s a detailed science to determine safe light levels for your eye, which needs to be well understood before illuminating your fundus. Candlelight is generally safe, but please consider these standards and don't look into bright lights for long periods of time.

Disposable camera lenses are relatively close in focal length to a human eye, so I 3D printed some parts for mounting it. To mimic structure on the fundus, so I glued thread, yarn, and a resistor (why not?) onto the back part of the test eye. There’s no internal fluid in my test eye like a human eye’s gel-like vitreous, but it’s good enough for testing out my eye-imaging prototypes. For a quick test, I illuminated the test eye and peaked in to see the threaded fundus I made.

3D print a few parts attached here. Careful handling the beamsplitter. Place it into the mount and lock it into place with a few metric bolts. Cut a dowel rod and place it into the handle. Then add a pilot hole into the dowel and screw in the head of the device into the handle. A correction lens needs to be added depending on the patient's refractive error. I had a weak negative lens to use, but it is easier for the person being imaged to just wear contacts.

With the device built, it's possible to view someone's fundus by candlelight. The alignment is very challenging and slow burn candles are essential. Drawing out the light path is helpful too (see the paper in the first image). The person being imaged needs to fixate at a distance. Start moving towards them until you begin to see a faint red glow coming from the eye. You will eventually begin to make out some blood vessels in the fundus. The field-of-view will be small and the quality won't be great, but it's still a view of the fundus by candlelight!

To get a video of this, I set up a mount with a dSLR and aligned myself using a monitor behind the device. The gif above isn't high quality, but you can just make out the vessels. I have a picture of the fundus using a direct ophthalmoscope that has an LED, much better quality.