Pre-build lumber quality checklist

Before you cut any wood, walk through this checklist with the lumber you’ve bought. The wind and post-bending analyses assume clear, sound lumber — knots, slope-of-grain, decay, and wane reduce strength significantly. The NDS values the wind calc references are for Select Structural grade; most lumber-yard “construction grade” 4x4 is actually #2 or worse.

If you find any of these issues on the lumber you’ve bought, return it to the lumber yard and pick different boards. Do not “work around” it by moving the cut to a different location — the structural path is well-defined and bad wood on that path means a real-world failure mode.

Check every board you plan to use

• No knots larger than 1” in the bottom 12” of the four 4x4 corner posts. The base of the post carries the bending moment (see analysis/post_bending.py). A knot at that location reduces the cross-section’s effective area and breaks the SF > 1.5 target.
• No knots larger than 1/3 of the board face anywhere along the post or rail length. Smaller knots (>1” but <1/3 of the face) are acceptable at mid-height where the moment is lower.
• No through-checks or splits. A crack running through the board face means the fibers are already separated; the bending moment will open it further. Tap the board on the ground — a dull thud indicates internal decay; a clear ring is sound.
• No wane (missing wood) on the load-bearing edges. A 4x4 with a rounded corner at the bottom is fine for appearance but not for a column under a 6 ft lever arm.
• No twist > 1/4” over the length of the board. Check by laying it on a flat surface. Twist compounds as the frame goes together; > 1/4” twist at the lumber yard is > 1” twist in the assembled frame.
• No cup > 1/8” over the width. Cup is OK in the bed skin boards (the cleats hold them flat) but not in the rails.
• Moisture content < 19% (KD-19 or kiln-dried). PT lumber is usually wetter at delivery. If it’s > 19%, let it acclimate in the build area for a week or sticker-stack it before installing. Wet PT will shrink and open joints as it dries.
• Stamp check — the grade stamp should be legible (“S-P-F”, “Douglas Fir”, “Hem-Fir”, etc.). If you can’t read it, the lumber yard can’t certify the grade.

Pressure-treated (PT) specific

• Use UC3B or UC4A rated for ground contact. UC2 (“above ground”) is wrong for the bed cleats and skids. The lumber yard should have a tag or stamp. If unsure, ask.
• Modern ACQ or copper azole treatment, not CCA. CCA (chromated copper arsenate) is banned for residential use since 2003 in the US. ACQ is the modern equivalent and is what you’ll get from Home Depot / Lowe’s.
• No “checker” surface treatment failures. PT should have a uniform greenish tint and the incising pattern should be consistent. Misshapen checker pattern = the treatment didn’t penetrate uniformly.

Cedar specific (for the skin boards)

• Heartwood > sapwood. Cedar’s rot resistance is in the heartwood. Sapwood (the lighter outer ring) will rot within 3-5 years when buried. Look for boards with mostly heartwood (the darker, aromatic inner wood).
• No wane on the face. Wane on a 1x6 skin board means less wood holding the cleat. The cleats are at most 24” o.c., so any board with wane in the middle 12” of its face is unusable.
• No “skip” (planer misses). Cedar is often rough-sawn at the lumber yard. Skip is cosmetic for the garden bed but matters where the cleat screws bite into the skin.

Stainless / hot-dip galvanized hardware check

• Screws are #10 or larger, ACQ-compatible. ACQ eats plain zinc-coated screws in 6-12 months. Use Simpson Strong-Tie Outdoor Accents or equivalent. Box of 1 lb is ~$8.
• Bolts are hot-dip galvanized or stainless. The hinge bolts, actuator mount, and panel clamps carry significant load. Don’t use electro-galvanized (zinc-flash) bolts in ACQ-treated lumber — they corrode.
• No mixing of galvanized and stainless. Galvanic corrosion between dissimilar metals is real; if you use stainless bolts in a galvanized bracket, the bracket corrodes faster. Pick one metal and stay with it.

Site prep

• Bed footprint level within 1” over 8 ft. Use a 4 ft level + straight edge. The 35° tilt assumption in the wind calc is for a level bed; a > 1” tilt in one direction adds a moment the calc didn’t budget for.
• No overhead wires or trees within 10 ft of the actuator’s max extension. At 35° tilt the panel + post top is ~7.5 ft up. A windstorm that drops a branch on the panel will destroy the actuator AND the panel.
• Southern exposure, no shade between 10 AM and 4 PM year-round. Use the SunAP (iPhone) or Sun Surveyor (any phone) to verify before you commit to the bed location.

Don’t

• Don’t use boards with heart-rot even if it’s “small” — it grows.
• Don’t use lumber marked “stud grade” for the structural members (posts, rails). Stud grade is graded for non-structural use; it’s not the same as #2 structural.
• Don’t assume “new from lumber yard” = “dry”. Most PT is wet at delivery; budget for a week of drying time.
• Don’t substitute 2x4 for 4x4 posts — the moment of inertia is 8x different (I = bh³/12; 3.5³ vs 1.5³ = 12.7 ratio). The post-bending analysis assumes 4x4.

When in doubt

Post-bending analysis (analysis/post_bending.py) says the 4×4 fails at 35° unbraced. The two remedies are:

• Upsize to 6×6 (SF 2.53 with the same loads)
• Add lateral bracing (square-cut gusset plates or off-the-shelf structural angle brackets)

If your lumber is borderline (knots at the base, lower-grade stamp), pick one of those remedies. The wind calc is a first-pass model with a 2.0 SF target; the PE review recommended in the README is the right move if you have any doubt about the lumber quality.