Busting Random-Orbit Sanding Myths

Questions abound about the proper use of the random-orbit sander: How fast should I move it? How hard should I press down? Do I start the sander on or off my workpiece? Rather than spout conventional wisdom, we challenged it in our shop and came up with several strategies you can use in yours. Here's how we tested these techniques. (To learn the results of our tests, pick up the October 2008 issue of WOOD magazine, and turn to page 60.)

Knowing the impact of load on the sander, and how it bears on both speed and motion, sheds light on some thorny finishing issues, including where to start the sander and how pressing down on the tool affects performance.

Here, we used a stopwatch and optical tachometer to test the ramp-up of a sander motor, both under no-load conditions (shown) and with force applied. The tachometer, which measures rotational speed, shoots a laser through a vent in the sander housing to read a reflective marker inside.

We used a similar method to test the brakes for the sanders' backing pads. Unchecked, the free-wheeling pads can whirl to wood-marring speeds. This makes off-the-board starts dicey for older, brakeless sanders.

Under normal load, a sander motor must overcome the friction between the abrasive and the wood. This resistance increases with the coarseness of the sanding disc, the roughness of the wood, the sander's weight, and the pressure applied by the operator's hand. For testing purposes, we defined "normal load" as the weight of the sander plus our plywood yoke, shown, which served as a proxy for the weight of a hand. We used fresh 100-grit sanding discs and newly planed maple to establish our baselines.

In this picture, the optical tachometer shows a sander, rated at a no-load speed of 12,000 orbits per minute, approaching its top speed under normal load.

Here, stain reveals the pattern left by an on-the-board sander start. In attempting to produce reliably clean starts, and reliably poor ones, we tried several techniques. We tested off-the-board starts with a running sander; we lowered the sander straight down; and we also practiced bringing it in for a landing. We worked on both clamped and unclamped work pieces, and screwed boards to the wall to see if vertical starts would create issues we hadn't seen on horizontal starts.

Does pressure on the sander help or hurt? And how much is too much? To find answers, we needed a means of controlling downward force. Our test apparatus (pictured) employed small dumbbell weights centered above the sander on a dowel. We used this rig both freehand, and on a stand, paired with a stock feeder. We visually evaluated the effects on the sanding surface and used instruments to gauge stress on the tool.

We employed an infrared thermometer to measure the heat thrown off by sanders and sanding discs under varying conditions and loads. Here, under no load, the thermometer puts the sander's air-cooled metal base modestly above room temperature. Our tests confirmed that under heavy load -- such as that experienced when one bears down on the sander -- components like discs and hook-and-loop fasteners become prone to overheating and early failure.

Disc speed is one thing, but how fast should you move the sander itself for the best results? To test the effects of different linear sanding speeds on surface quality, one of our experiments utilized a stock feeder to move boards beneath a random orbit sander. We double-checked the feed-rate settings by using our tach to count wheel revolutions. (A small piece of reflective tape for this purpose is visible on the center wheel in the photo.) Knowing the RPMs and diameter of the wheel allowed us to figure inches per second of travel. We made additional observations using a stopwatch and rules.

How fine must you sand to create a professional finish? Answering this question didn't require any special test equipment, just stacks and stacks of new sanding discs, an industrial-size carton of dust masks, and long days in the finishing room. We focused our efforts on the most common wipe- and brush-on consumer finishes, including polyurethanes and poly-acrylics, and to a lesser extent, lacquers and oil blends.

We worked with side (raked) lighting, which is invaluable for spotting sanding and finishing flaws. As the accompanying photos illustrate, a fairly nasty sander swirl, obvious when lit from the side (bottom), can all but disappear with a small change in angle (top). We later applied a pre-stain conditioner and oil-based stain to our test boards, which made the sanding marks easier to observe and photograph.

As with most of our tests, we relied primarily on freshly planed hard maple and eastern white pine boards with similar moisture contents. We sprinkled in a few other soft- and hardwoods for good measure.

To help ensure that differences we observed in samples were caused by grit changes and other sanding variables we wanted to test -- and not disparities in grain patterns, color or density between boards -- we made side-by-side comparisons using either bookmatched lumber or adjoining sections cut from the same board.

We resawed the pine samples (top pair) from one board. The bookmatched halves differ only in how finely each was sanded. We crosscut the cherry samples (bottom pair) from a single board and, likewise, sanded them to different levels. Both examples show that roughly sanded wood (right top and bottom) holds more stain or finish, and therefore is apt to be darker, than fine-sanded wood (left top and bottom).

Some folks argue that random-orbit sanders are too aggressive for final finish work. So we put the tools to the test -- both for between-coat scuffing and for rubbing out final coats. We worked with stearate-coated aluminum oxide sanding discs and newer mesh abrasives, in grits from 220 to 1500. We focused mainly on polyurethane finishes, conducting trials at full speed and several slower settings.

This picture shows a cherry board finished with a water-based poly-acrylic. In this case, we sanded after each coat at full speed with a 320-grit abrasive mesh disc (also shown). (We marked the sample number, 119, on the edge of the board for reference.)

We analyzed scratch patterns after each test. Normally, a random-orbit sander cuts connected ringlets, like those pictured, which are blended away as the machine works and as passes overlap. A well-maintained sander, used with sound technique and fresh abrasives, leaves a path that is thoroughly and evenly scuffed, with scratches of uniform depth. The existence of unblended patterns can signify that you are moving the sander too quickly, applying pressure inconsistently, or using discs with uneven size abrasive grit. In any case, they mean you aren't done sanding.

Deep swirls, uneven scratch patterns and visible scores -- like the marks shown here -- signal something's gone awry. We produced these by applying extra pressure to the sander, which, among other things, hampers the normal movement of the free-spinning backing pad. Other potential causes of swirls include using a sander with worn or damaged parts, and sanding wood that is too wet.

Now you know how we conducted our tests. To find out how our test results can help you finish your projects faster and better, pick up the October 2008 issue of WOOD magazine, and turn to page 60.