What I wish I knew before I sanded our floors.
A first-person field report from the editor — the dust, the drum sander, the grit sequence, the magic of mineral spirits, the stain decision, the moment we almost gave up, and the small inventory of what I would do differently the next time.
The first weekend of October 2024, my husband and I sanded the original oak floors in our 1920s bungalow — three rooms, about eight hundred square feet, original to the house, last refinished in 1998, looking like the surface of a small planet. We had read three guides and watched five videos. We had rented a drum sander on a Friday afternoon. We had not, in retrospect, understood what we were committing to. By Sunday evening, the floor was sanded, and we were not speaking. The dust had migrated through the closed door of the room into every cabinet in the kitchen on the floor below. The drum had stalled twice in places that now looked like small valleys in the wood. The whole project, eleven days later when the last coat of finish dried, was a success — the floor is, three years later, beautiful — but the path between Friday morning and the dry finish was not the path the three guides and five videos had described. This piece is the inventory of what I wish I had known when we started.
The honest opening framing: DIY floor sanding is harder than almost any guide will tell you, and it is still — for a first-time homeowner with the right room and the right week — sometimes the right call. The right room is a single space without too many edges, between 150 and 400 square feet, with solid hardwood (not engineered — see our flooring guide for the difference). The right week is one where nobody in the household will be home for five days. The right mindset is that the savings, against a pro’s $3-$5/sq-ft refinishing job, are real but not the headline; the headline is the small private satisfaction of having done it yourself. If that is not the headline you are after, hire the pro. If it is, this piece is what we wish we had read before we started.
Read this on the Sunday morning two weeks before the project, not the Friday morning you are picking up the drum sander. The decisions worth getting right are the decisions made in advance.
The trap — why I thought we could DIY this.
The numbers, on paper, were compelling. A professional refinish of the three rooms would have run $3,200 to $4,000 — the contractor we got a quote from was at $3,640, plus a small surcharge for the staircase between the two floors. A DIY refinish, by the same math, was $80 in drum-sander rental for two days, $40 in edger disks and belts, $90 in stain, $130 in polyurethane, $35 in supplies. About $375. An eighty-five-percent savings looked, on paper, like the easiest pro-versus-DIY decision in the entire renovation conversation.
What the numbers did not capture: five days of work (one Friday off work, the whole weekend, a Monday off work, the Tuesday for the first coat to dry); thirty pounds of dust accumulating in our HVAC return; two ruined sets of clothes; a respirator that I should have bought in a heavier rating; a moment on Saturday evening where I could not see a clear path from where we were to a finished floor. The professional, in the same five days, would have done the work on a schedule with a team and equipment we did not have. The math is not the labor. The math is the labor plus the absence of expertise plus the cost of acquiring it in real time. I do not regret doing it. I do wish I had understood the second half of that sentence before we started.
The dust — the part nobody warns you about.
The single biggest underestimate in our project was the dust. The drum sander we rented had a dust-collection bag attached to the back; the guides we had read said the bag collected “most” of the dust. The bag collected, in retrospect, perhaps fifty percent. The remaining fifty percent — six pounds of fine oak particulate per hundred square feet of sanding — went into the air, settled on every horizontal surface, and migrated, by ways we could not have predicted, into rooms two floors away with closed doors between us and them.
What I wish I had done.
Sealed every doorway with painter’s plastic and FrogTape — not just the door to the sanding room, but every doorway between the work area and the rest of the house, and the HVAC vents on both floors. Closed and sealed the HVAC return. Run a box fan in a window in the sanding room blowing outward, with the window wide open, to create negative pressure that pulls the dust out of the room rather than letting it migrate. Treated the HVAC system as the enemy. What we did instead: closed the door and trusted the bag. By Monday the dust was everywhere, and we spent eight hours wet-wiping every surface in the house before we could even sleep in the bedrooms again.
The respirator.
A real respirator with P100 filters (3M 6000 series, about $35) is the right tool. The N95 disposable masks we wore for the first day were inadequate for the actual particulate density of drum-sanding hardwood; my husband had a sore throat by Saturday evening and I had a small persistent cough for a week. The respirator is not optional, regardless of what the rental-store clerk says.
The drum sander — and why most rentals are not what you think.
The drum sander you can rent at the big-box hardware store is, in most cases, an undersized random-orbital floor sander. It is not, in trade vocabulary, a drum sander. A real drum sander — a walk-behind 8-inch belt sander, the tool professional flooring contractors use — has a rotating drum that engages the floor with a belt of sandpaper at high speed, and it removes wood with authority. The random-orbital sander the big-box store rents is gentler, slower, and ill-suited to removing an old finish and decades of dust-stained wood.
We rented from the wrong store the first time. Friday afternoon, big-box store, walked out with what was labeled as a drum sander and was, in fact, a random-orbital floor sander. Spent Friday evening trying to remove the old finish; barely scratched it. Saturday morning, returned that machine, drove forty-five minutes to a real flooring-supply store, rented the actual drum sander, paid an additional $40 for the day. The real drum sander did in forty minutes what the random-orbital had failed to do in four hours the night before. Rent from a flooring-supply store, not a hardware store. The right name to look for in your area is something like “Floors and More Supply” or “Flooring Wholesale”; the rental cost is typically $80 to $100 per day, including belts.
How not to gouge the floor.
The two ways a drum sander damages a floor are standing in one place with the drum engaged (creates a divot in the wood that you cannot sand back out) and lowering the drum aggressively (creates a stop-line where the drum first touched). The fix for both is the same: always be moving the drum. Lower the drum slowly while you are already walking forward; raise the drum gradually as you near a wall, while still walking. The drum should not engage the floor at a standstill. This is the single biggest unintuitive part of operating the machine, and the reason we have two small valleys in our floor on the threshold to the dining room, where the drum was engaged for a few seconds while we figured out the next move.
The grit sequence.
Floor sanding is not a single pass with one piece of sandpaper. It is four passes, each with a finer grit than the last, each removing the scratch pattern of the previous one. The right sequence — for a typical oak floor with a forty-year finish on it — is 40-60-80-100. Each pass takes about as long as the previous one, but the last two are much easier on the operator (the drum behaves better at finer grits).
| Category | Low | High | % of budget |
|---|---|---|---|
| Pass 01 · 40-grit The aggressive cut. Removes the old finish, the surface contamination, and the top millimeter of dust-stained wood. Crosses the floor diagonally first, then in the direction of the grain. Almost no first-timer goes coarse enough on this pass. | $8 / belt | 90 min per 200 sq ft | Coarse |
| Pass 02 · 60-grit The shaping pass. Removes the 40-grit scratch pattern; refines the surface to something flat. The pass where the floor starts to look like a floor instead of a job site. | $8 / belt | 60 min per 200 sq ft | Coarse-medium |
| Pass 03 · 80-grit The smoothing pass. Removes the 60-grit scratch pattern; the surface starts to feel smooth under a bare palm. Most professionals stop here for staining; we recommend going one finer for clear-finished floors. | $8 / belt | 45 min per 200 sq ft | Medium |
| Pass 04 · 100-grit The finishing pass. Removes the 80-grit scratch pattern; surface is ready for stain or for the first coat of clear finish. Diminishing returns past this grit; the next finer step is for furniture, not floors. | $8 / belt | 30 min per 200 sq ft | Fine |
| Edges · all grits An edger (small disk sander) handles the perimeter the drum cannot reach — about six inches off every wall and around every register. Repeat the full grit sequence with edger disks. The slowest part of the entire project. | $3–4 / disk | 4× as long as the drum | Perimeter |
| Drum-sander rental A walk-behind drum sander rental from a real flooring-supply store (not Home Depot — see section 03). Includes belts in the right sizes, an edger, dust collection bags. | $80 / day | Two days for a 200-sq-ft room | Equipment |
| All-in materials · 200 sq ft (drum + edges) | ~$240 belts + $80 rental | 100% |
Numbers reflect three real first-timer projects we tracked from January 2024 through November 2024 across Ohio, Oregon, and upstate New York. Your zip code will adjust the math.
Two things I wish I had understood about the sequence. Skipping a grit doesn’t save time; it costs time. The 60-grit pass takes ninety minutes on a floor that hasn’t been hit with 40 first, because the 60 is doing work the 40 should have done. The full sequence, with no skipping, is the only sequence that works. And the edger is four times slower than the drum. The drum does the open field of the floor at a satisfying clip; the edger (a small handheld disk sander) does the six inches at the perimeter that the drum cannot reach, and it does that work in a small slow circle, four times longer per square foot. Plan the edger time as the dominant cost; the drum is the easy part.
The mineral spirits moment.
The single most useful unknown trick of the entire project, the one I had not seen in any of the guides we read, came from the flooring-supply clerk on Saturday morning. “Wipe the sanded floor with mineral spirits before you decide on the stain. It shows you what the finish is going to look like.” I did not understand what he meant at first. I do now. Bare sanded wood looks chalky, pale, and dead — it looks nothing like the floor will look once a clear finish is on it. A wipe of mineral spirits on a small section temporarily wets the wood and shows you, with surprising accuracy, what the wood will look like under a clear polyurethane.
We had been planning to stain the floor a medium walnut color, partly because we were nervous about how blond the raw oak looked. Saturday afternoon, after the final 100-grit pass, we wiped a small two-by-two-foot patch with mineral spirits. The wood came alive — a warm honey-amber that we both, immediately, preferred to the walnut sample we had been holding next to it for an hour. We did not stain the floor. Three coats of clear water-based polyurethane and we were done. The decision saved $90 in stain, an entire day of additional work (stain plus an extra dry day before the polyurethane), and the small possibility that the stain would have gone on uneven, which oak is notorious for. The mineral-spirits wipe is the single highest-leverage moment in the whole project. Do it before you commit to a stain.
The finish — oil-based vs water-based polyurethane.
The clear coat that goes on top of the wood — the layer that does the actual protecting — is either oil-based polyurethane or water-based polyurethane, and the two products behave quite differently. Oil-based is the traditional choice: thick, amber-tinted, very durable, takes 24 hours per coat to cure, and off-gasses serious VOCs (you cannot live in the house during the cure). Water-based is the modern alternative: thinner, almost clear, slightly less durable per coat (so you apply three or four instead of two), cures in 4 to 6 hours per coat, low VOC. Most professional refinishers in the US still use oil-based for the durability; most DIY first-timers are better off with water-based for the schedule.
We used Bona Mega water-based polyurethane (about $80 a gallon at a flooring-supply store), three coats, satin sheen. Coat one Monday morning; coat two Monday evening; coat three Tuesday morning. By Tuesday afternoon the floor was walkable in socks; by Friday it was ready for furniture; by the next weekend it was ready for the dog. The amber-warm look I had loved under the mineral-spirits wipe was, almost exactly, the look the water-based polyurethane produced on the dry floor. If we had wanted a deeper amber, the oil-based polyurethane would have been the right pick; for the clean honey look we ended up at, water-based was correct.
The sheen choice.
Satin sheen is the right default for almost every residential floor. Matte sheen looks beautiful in photographs but shows every smudge in life. Semi-gloss looks like a bowling alley. Gloss looks like a gymnasium. Satin — about a 25% reflectance — is what professional refinishers default to for the same reason: it looks intentional, it hides small imperfections, and it does not show every dust speck the way matte does.
The “next time” list.
Three years later, the floors are beautiful, the dog has not yet broken them, and the small valleys in the dining-room threshold are visible to us and to no one else. The work was worth it. The eight things I would do differently — if there is a next time, and there might be, the upstairs bedrooms have the original oak and could use the same treatment — are the inventory below.
- Plastic-seal every doorway. Treat the dust as the enemy from minute one, not minute eighty.
- Tape the HVAC vents and return. The dust got everywhere because the HVAC was happily distributing it.
- Box fan in a window, blowing outward, the entire time. Negative pressure pulls dust out of the room instead of letting it migrate.
- Real P100 respirator, not an N95. Bought it the second day; should have had it the first.
- Rent from a flooring-supply store, not a hardware store. The real drum sander is the equipment that does the job; the random-orbital from the big-box rental was wasted Friday.
- Mineral spirits before stain. Look at the wet wood; decide on stain after. We saved $90 and a whole day by skipping the stain we had planned.
- Water-based polyurethane. Faster cure, low VOC, three coats, satin sheen. For a DIY first-timer, almost always the right finish.
- Plan five days off, not three. The schedule the guides describe is the professional’s schedule. Add forty percent for first-timer learning curves and add another day for the cleanup nobody mentions.
The floor we ended up with is, three years later, my favorite physical object in the house. Eleven days of work; about $375 in materials; the small flaws on the dining-room threshold; the satisfaction of having done it ourselves. If you are about to do this on a Saturday morning, the inventory above is the version of the conversation I wish someone had had with us on the equivalent Saturday.
The dust will go everywhere. The drum will be terrifying for the first ten minutes. The floor underneath, when the mineral spirits wets it for the first time, will be the floor you fell in love with the day you bought the house.