Torque wrench calibration

I have been using a Toledo torque wrench on the elevator, wjngs, and shortly to rebolt the spar cap. It came with a calibration certificate, but I thought I would verify its accuracy.

When I torqued the elevator bolts, it was set to about 38 in-lbs by eye to torque the elevator bolts, naïvely interpolating between the 35.4 in-lb and 53.1 in-lb markings rather than using the vernier. I have left it at this setting. The scale advances one step per half turn, so 53.1-35.4 = 17.7 in-lb per 1/2 turn The vernier itself is oriented to the N-m scale, with 1/2 turn advancing 2 Nm (which is 17.7 in-lb). The scale was advanced 1.38Nm past the 35.4 in-lb marking, so the torque setting was in fact 35.4 + (1.38/2.0)*17.7 = 47.6 in-lb. This is nominally too high for AN3 bolts, but the following measurments bring the actual torque back to 40 in-lb.

Note that when I refer to, say, the 53.1 in-lb torque setting, I am now lining up the vernier to zero, with the barrel nearest the 53.1 in-lb marking on the barrel. In fact the lineup between the edge of the barrel and the 53.1 in-lb division is imprecise. The torque would be set slightly higher if I made an adjustment based on the alignment of the collar with the markings on the barrel (as I did with the elevator bolts), but this is hard to standardize owing to the width of the imprints on the barrel. So in the following I am taking the former approach and using the vernier to standardise on the "zero" of the barrel markings

Vertical-arm measurement scheme:

A direct check was made using a spring balance and arm, ensuring the wrench clicked when the arm was vertical (so that the torque due to the weight of the arm did not affect the result).

Arm length (from socket centre to hook eyelet) 700mm = 27.56"

Checking the setting used to torque the elevator bolts:

Setting = 47.6 in-lb Force 0.60 kg => torque = 36.4 in-lb

Further tests:

Setting = 141.6 in-lb Force 1.97 kg => torque = 120 in-lb

Setting = 123.9 in-lb Force 1.70 kg => torque = 103 in-lb

Setting = 106.2 in-lb Force 1.52 kg => torque = 92.3 in-lb

Setting = 88.5 in-lb Force 1.25 kg => torque = 75.9 in-lb

Setting = 70.8 in-lb Force 0.95 kg => torque = 57.7 in-lb

Setting = 53.1 in-lb Force 0.67 kg => torque = 40.7 in-lb

Setting = 35.4 in-lb Force 0.45 kg => torque = 27.1 in-lb

Setting = 17.7 in-lb Force 0.20 kg => torque = 12.1 in-lb

This level of error is surprising, given the torque wrench was provided with a calibration certificate, which showed a much greater accuracy.

The Pesola spring gauge was checked with a 1000g and 5000g weight, and was found to indicate an error within 1%

Horizontal-arm measurement scheme:

These tests are repeated with the scale vertical and the arm horizontal. The scale tare setting is adjusted to zero out the weight of the arm before pulling it upwards against the resistance of the wrench.

Setting = 141.6 in-lb Force 2.00 kg => torque = 122 in-lb close to before

Setting = 123.9 in-lb Force 1.75 kg => torque = 106 in-lb close to before

Setting = 106.2 in-lb Force 1.54 kg => torque = 93.4 in-lb close to before

Setting = 88.5 in-lb Force 1.25 kg => torque = 75.9 in-lb same as before

Setting = 70.8 in-lb Force 0.95 kg => torque = 57.7 in-lb same as before

Setting = 53.1 in-lb Force 0.68 kg => torque = 41.3 in-lb close to before

Setting = 35.4 in-lb Force 0.53 kg => torque = 32.2 in-lb

Setting = 17.7 in-lb Force 0.30 kg => torque = 18.1 in-lb

Apart from some differences at the bottom end, it is clear that the orientation of the scale and arm (once its own torque due to its weight was zeroed out) does not affect the general conclusion: the torque wrench is subject to significant error.

Rechecking the setting used to torque the elevator bolts:

Setting = 47.6 in-lb Force 0.66 kg => torque = 40.0 in-lb

With the vertical arm measurement scheme, a torque of 36.4 in-lb was obtained. I think we are right on target near (or maybe slightly above - further tests suggest 42.5 in-lb) the max torque setting for AN3 bolts (with fresh nylon-locking nuts) and very close to the revised TAF recommendation of 38 in-lb. The manual figure of 25 in-lb was rather low, almost "loose".

So revisiting torqueing these bolts is probably not warranted, and I can proceed with rebolting the spar caps.

Do I trust the calibration table? Maybe not. I think I will trust my measurements instead.

Vertical-arm measurement scheme: Error due to arm not being quite vertical

With the arm horizontal the spring-balance force (with it zeroed under no load) is 0.45kg - this is an effective torque of 27.3 in-lb

With the arm at angle theta from the vertical this effective torque is 27.3 * sin(theta)

If the torque wrench is arranged to click within 5 degrees of the vertical, the effective torque is at most 27.3 *sin (5 degrees) = 2.4 in-lb

This can be regarded as the uncertainty for the vertical beam measurements

I'll repeat some tests again, under better controlled conditions.

Setting for torquing elevator nuts Force 0.74 kg => torque = 44.9 in-lb

Setting = 212.4 in-lb Force 3.30 kg => torque = 200 in-lb

Setting = 196.7 in-lb Force 3.08 kg => torque = 187 in-lb

Setting = 177.0 in-lb Force 2.81 kg => torque = 171 in-lb

Setting = 159.3 in-lb Force 2.52 kg => torque = 153 in-lb

Setting = 141.6 in-lb Force 2.20 kg => torque = 134 in-lb

Setting = 123.9 in-lb Force 1.95 kg => torque = 118 in-lb

Setting = 106.2 in-lb Force 1.68 kg => torque = 102 in-lb

Setting = 88.5 in-lb Force 1.40 kg => torque = 85.0 in-lb

Setting = 70.8 in-lb Force 1.16 kg => torque = 70.5 in-lb

Setting = 53.1 in-lb Force 0.85 kg => torque = 51.6 in-lb

Setting = 35.4 in-lb Force 0.54 kg => torque = 33.0 in-lb

Setting = 17.7 in-lb Force 0.31 kg => torque = 18.8 in-lb

Repeating the last 4 measurements with 350mm arm (using hole drilled at the midpoint)

Setting = 70.8 in-lb Force 2.16 kg => torque = 65.6 in-lb

Setting = 53.1 in-lb Force 1.66 kg => torque = 50.4 in-lb

Setting = 35.4 in-lb Force 1.15 kg => torque = 34.9 in-lb

Setting = 17.7 in-lb Force 0.69 kg => torque = 21.0 in-lb

Since the discrepancy above is often comparable to the uncertainty to to the effective torque, and other evident factors I have now a bit more confidence in the calibration of the torque wrench, but will regard these measurements as a cross-check. I may revisit the elevator nuts.