Harry Bloomfield wrote:

And I can visualize how to mechanically clamp a bolt in a vise to hang the
torque wrench on - but then - how do you calibrate the two types?

I can imagine, when building them that they simply check the
calibration on a pass / fail basis. The only way I can think to adjust
them is via shims under the spring, to increase the tension.

From what I saw on YouTube, there's an SAE standard for calibration of
torque wrenches, which, as I remember from the video of a day or two ago,
comprises three sets of five tests, summarized (from memory) as:

1. Unmeasured five full-torque applications (e.g., 150 foot pounds)
2. Measured, five 20% torque applications (e.g., 30 foot pounds)
3. Measured, five 60% applications (e.g., 90 foot pounds)

Average it all out and you get your percentage error.

If the error is worse than what you desire, you adjust the calibration of
the torque wrench (where some have hex adjustments of spring tension,
others have twist adjustments, others have bend adjustments, etc.).