Revised flashlight test

Posted May 6, 2004

Introduction

I would like to propose a simpler method to conduct initial testing on a rough-ground telescope mirror. I know that many other amateur telescope makers must have stumbled across this method of testing, although I have yet to see it documented anywhere.

Important information: I am, as of this writing, a rank novice at mirror grinding. I've put about 1.5 hours into my 6" mirror so far. I may be wrong about how this test works, but I'm always interested in contributing where I can.

Background

Described in many books and on many websites is a "flashlight test" that can be used to determine the approximate focal length of a mirror. Hold a flashlight next to your head and move around until you can see its reflection in your wetted, rough-ground mirror. When you find the distance at which you can no longer see the flashlight's reflection moving, but can only get the mirror to appear "bright" or "dim", that's the radius of curvature of the mirror.

However, this test is difficult to accomplish for the novice, for it's hard to judge exactly when you're really seeing the mirror go bright or dark (as opposed to merely being close, and seeing the light move on and off the mirror a little bit). Even then, it's difficult to get a measuring tape near enough your face and the mirror that you can make accurate measurements.

This is a modification of that test. Noticing that I could see a bright spot on my glasses as I tried to find the correct length, I decided to try putting a piece of paper around the flashlight, and just find the distance at which the image was smallest.

This worked better than I had expected. Thus I present it here.

The method

First, construct a viewing screen. I took an 8.5x11" piece of copy paper, and taped it to a piece of cardboard. I then cut a hole in the cardboard and paper such that a small LED flashlight would be held fairly tightly in the hole. It'll wear out before too long, but it's also easy to replace. Try to ensure that the "plane" of the flashlight is equal with the plane of the paper, for the most accurate results.

Once you have your viewing screen, lay out your mirror and measuring tape in a dimly lit room. I used a plastic drinking cup to hold up the mirror -- cheap, easily available, and won't damage the mirror. You can use a cup, or construct a stand, anything that'll leave the mirror aiming in about the right direction, along the measuring tape. Align the end of the tape so that it's directly underneath the front edge of the mirror. Wet the mirror with a mixture of water and dish detergent, which will help it coat the mirror and keep it from evaporating as quickly.

Referencing figure 1, move your viewing screen in the B and C directions until you can see some kind of reflection on the screen. Offset it from the actual flashlight, so you can see a good image. Now, move the screen-and-flashlight in the A direction (toward and away from the mirror) until you see the image getting smaller.

When you see the image (D) of the flashlight get to its smallest point, the flashlight/screen combo is at the radius of curvature of the mirror. It'll never be a perfect reflection as long as you're dealing with a rough-ground mirror. The smallest point should be about the same size as the light-emitting element of the flashlight -- the reflector on an incandescent flashlight, or the LED(s) on an LED flashlight.

Note, as well as possible, where on the measuring tape the plane of the screen falls. Write that down in your testing/grinding log, and divide it by 2 to get the approximate focal length of the mirror.

Caveats

Like any test of this nature, this one is imprecise. I believe, based on my experience so far, that this test has a greater chance of being repeatably accurate by amateur telescope makers with no experience whatsoever. I know when an image goes from being small to being larger and vice versa. I don't really know what the books mean when they talk about the mirror "going bright" and "going dim" without any movement being visible.

I suspect that the reasonable accuracy of this test is around 2-3" for the radius of curvature. I definitely think it's more accurate than trying to view the mirror surface directly. Since the mirror must be wetted with water or something similar in order to make it reflective, the test will never be accurate -- the water will always pool at the bottom of the mirror, ruining the apparent curve of the glass in that area.

Finally, as I said above, I've only been doing the ATM thing for a week or so at this point. I've read a great deal about it, but my thinking may still be inaccurate on a number of subjects. I may look back on this in a month/year/decade and laugh at my naivete, but I believe it to be a good test now.

Created by Ian Johnston. Questions? Please mail me at reaper at obairlann dot net.