Bandsaw Blade Tension: Why the Gauge Isn't an Instrument (and How to Measure Stretch)

Yo, let me tell you about the sound my apartment makes at 11pm when I'm resawing.

Nothing. That's the point. Second bedroom, thin walls, neighbor on the other side who has been unbelievably cool about all of this. So I run a 14" saw slow and I baby my blades, because a blade that snaps at 11pm is a conversation I don't want to have through drywall.

Which is how I ended up staring at that little tension window on the upper housing, thinking: what is this thing actually telling me?

Turns out the honest answer is on record, from the people who built the saws.

The Gauge Is a Spring, Wearing a Costume

Grizzly says it out loud in the G0555 manual. You tension by the scale, then you run the saw, and then the manual tells you that what you just did "is only an approximate tension" and here's a fine-tuning procedure to fix it. Their own troubleshooting table has a row whose symptom is, verbatim, "Blade tension scale is grossly inaccurate." Cause: the spring lost its spring, because you left tension on it or over-tensioned it.

Laguna's 14BX manual is blunter. "Tension indicators are designed to indicate the compression of a spring." And then: "The tension scale is a general reference and not a rule." Use it "only as a general guide."

Rikon skips the pretense entirely. The 10-306 manual gives you a "General Rule" — blade should deflect about ¼" under a finger — and doesn't reference a tension scale in that procedure at all.

Jonathan Sevy nails why none of it can be trusted even when the spring is fine: manufacturers don't label the gauge in force. They just put marks where the adjustment should sit for different blade widths. So it's a spring-compression indicator with blade-width stickers on it. WEN's page for the BA1487 sells you that same window as a feature.

The Number You're Chasing Belongs to the Blade

Here's the equation, and you can check it on your phone.

Tension (lb) = target PSI × blade width (in) × blade thickness (in)

Tension is a stress. Pounds per square inch of blade cross-section. Forrest Manufacturing works the example: a ½" × .025" blade needs 312.5 lb of pull for 25,000 PSI, and a 1" × .025" blade needs 625 lb for the same PSI, because you doubled the area. SawBlade.com publishes the identical pair of numbers. Two sources, same arithmetic.

Now the factor almost everybody misses. The blade runs as two spans between the wheels. Sevy works it: a .032" × ½" blade at 15,000 PSI wants 240 lb of blade tension, so the spring has to carry 480 lb. Quarter ton-ish across a frame you can pick up with a hand truck.

And what number are you even aiming at? Depends who you ask:

Source	Target
Forrest	15,000–25,000 PSI, carbon toothed
SawBlade.com	15,000–20,000 PSI carbon; up to 35,000 bi-metal
EZtension	a flat 15,000 PSI
Timber Wolf ½" silicon steel	10,000–13,000 PSI

That Timber Wolf row is not an error. Their band mill catalog states the blades "run purposely with 35% to 50% less tension than our competitors carbon blades," which kills over 70% of the rotation fatigue effect and needs 20% less horsepower. Set a Timber Wolf ½" to a generic 15,000 and you've blown past its published max. Forrest, meanwhile, says nothing gets tensioned past 35,000 PSI, and recommends the lowest tension that still cuts clean.

There is no single correct number to calibrate a scale against. That's the whole game.

The Part Where the Fantasy Dies

Matthias Wandel's method is to clamp a caliper across a 4" span of blade. His worked example: a span that opens 0.002" is a strain of 0.0005, which times 29,000,000 PSI equals 14,500 PSI.

Then the number that should stop you cold. On a lever rig he pulled a blade to 360 lb before slip. On the actual 14" saw he got to 65 lb — and at 65 lb, the tension spring had started to bend. He describes the cheap 14" scale as suggesting tensions on the low side.

Do the arithmetic against Sevy's factor of two. A ½" × .025" blade at 15,000 PSI wants 187.5 lb, meaning 375 lb across the frame. A spring yielding at 65 lb isn't in the same ZIP code.

So for a lot of us the real question isn't "is my gauge off." It's "can this frame reach the number at all." And I'll say the unglamorous thing: no source I have publishes a controlled test of consumer saw scales against a calibrated meter with per-saw error figures. Anyone handing you "14" saws read X% low" is handing you a vibe.

Nobody Gets to Sell You One Correction Factor

Best evidence here isn't from the woodworking press. Kirbach and Bonac, at Environment Canada's Western Forest Products Laboratory, published in Wood and Fiber in 1978 on a production bandmill with a hydraulic strain system. They checked the manometer-type strain indicator against strain gauges and found its accuracy "vary considerably with tension-stress level," so every reading "had to be corrected." Their conclusion recommends more accurate stress-measuring devices.

Read that twice. The error varies with level. You cannot learn your saw reads 20% low once and apply it forever.

Same paper gives you the physics of flutter: lateral natural frequency drops with longer span and higher blade speed, rises with tension, and span length dominates. Which means the flutter point belongs to your saw, your blade, your speed. Not a universal PSI.

What I'd Actually Do in a Small Shop

Measure stretch. That's what the $300 Lenox 62126 does — clamp a fixed arm, read a dial indicator — and it's what Simonds' magnetic gauge does. Same physics, cheaper hardware.

PSI per 0.001" of stretch = E × 0.001 ÷ gauge length

John TenEyck built a 12" gauge from ¾" hardwood, ¼-20 bolts and thumbwheels, and a 0.001" dial gauge, for under $25. David Gunter 3D-printed a 6" version around a Harbor Freight dial gauge — $10–15 for the gauge, under $20 hardware — against commercial gauges he prices at $200 to $800.

Go long. At E = 29,000,000, a 12" span is about 2,417 PSI per 0.001"; a 6" span is about 4,833. Misread one division on the short gauge and you're off ~4,800 PSI — a third of a 15,000 PSI target. TenEyck's published 2,500 comes from using 30 × 10⁶; Gunter's 5,000 is 4,833 rounded up. Use 29 × 10⁶, per MW Components' 1018 data sheet.

Fair warning: Laguna's manual claims different blade steels stretch at different rates, so a gauge calibrated on one brand won't read right on another. The counterpoint is that modulus is stiffness, not strength — 1018 shows 29,000 ksi modulus alongside 63,800 PSI ultimate — and modulus barely moves across carbon steels. Nobody I can cite tested it directly. Pick your side knowing that.

You Found It, You Remember It

The flutter test finds a floor honestly. Timber Wolf's version: pull the guides and table insert, close the covers, run 30 seconds at speed, back tension off until you see a constant flutter — roughly 3/16" side to side on a ¼" blade in a vertical saw — call that zero, raise tension until flutter drops to about 1/16", then add 10% with ⅛ to ¼ turn.

How much you add above the floor is genuinely unsettled. Grizzly says one quarter turn past where flutter stops. Laguna says one full turn. Timber Wolf says the 1/16" point plus ⅛–¼. Those are materially different amounts of stretch, and I'm not going to pretend otherwise.

But step 12 is the payoff, and both Timber Wolf and Grizzly land on it: note where the indicator sits now. Grizzly's exact instruction is to use that reading as a guide for that blade in future — while warning not to lean on it long-term, because the blade stretches with use. Timber Wolf says the same and adds that thermal growth in the first few minutes of sawing may want another 5%.

That's the correction I can publish. Your gauge is not an instrument. It's a bookmark. Find the setting by measuring stretch over the longest span you can clamp, or by flutter, then let the scale remember where you were.

And detension when you're done. Suffolk says back off 8, 9 or 10 full turns and add the same back next session. Grizzly names spring set loss as the thing that wrecks the scale. Timber Wolf says left-on tension flattens the tire crown, and a crown worn to half your blade width means new tires — and until you swap them, the tension you need goes up. USDA's Forest Products Laboratory lists neglecting to relieve wheel strain when idle among documented causes of bandsaw cracking.

Cheap habit. Free. Buy the $25 of hardwood and a dial gauge instead of the meter.