How the calculator works

The Savarez Selector was the string calculator Savarez distributed on CD-ROM between 1998 and 2000: from a string's frequency, vibrating length and tension, it worked out its theoretical gauge in gut, nylon or fluorocarbon (KF) if plain, or its linear weight (PUL) if wound, and looked up the closest reference in the house catalogue. It stopped being distributed years ago; this page recovers and documents it with demonstrable fidelity.

This calculator reproduces that computation for the instruments in the original catalogue — or for a custom one — with live recalculation: pick an instrument, tuning, pitch standard and vibrating length, and every string instantly gets its gauge in the three plain materials or its PUL if wound, with the closest catalogue references. Tension is typed string by string, in its own cell, and there is an adjustment that raises or lowers all of them at once. There's no material to choose and nothing to click.

These are the two formulas it applies, worked out with a fixed example: a six-string classical guitar, 65 cm vibrating length and 7.5 kg per string, taking the 3rd string (G, 196 Hz) as the plain one and the 6th (E, 82.4 Hz) as the wound one. Those are the example's own values, not the calculator's 60 cm and 3 kg starting point: at 65 cm and 7.5 kg that 3rd string's theoretical gauge falls in the KF coefficient's second tier — the tier where this site diverges from the original's fixed coefficient (see Methodology) — while at the starting values it would fall in the first tier, where the two agree and there would be nothing to show. The example does not change when you pick another instrument: it is the results table that answers your settings.

How this was calculated

Plain

d = \frac{c \cdot \sqrt{1000} \cdot 100 \cdot \sqrt{T}}{f \cdot L}

Wound

PUL = \frac{245000000 \cdot T}{f^{2} \cdot L^{2}} \cdot 10

Original formula: d = c · √1000 · 100 · √T / (f · L) ; PUL = 245 000 000 · T / (f² · L²) · 10

Worked example:

Plain · SOL (G) 196.0 Hz · KF, KF tier 2, 100 ≤ d < 150 (c = 156.9767): d = 156.9767 · √1000 · 100 · √7.50 / (196.0 · 65.0) = 106.7 %mm (1.067 mm)
Plain · SOL (G) 196.0 Hz · Nylon (c = 170): d = 170 · √1000 · 100 · √7.50 / (196.0 · 65.0) = 115.6 %mm (1.156 mm)
Plain · SOL (G) 196.0 Hz · Gut (c = 155): d = 155 · √1000 · 100 · √7.50 / (196.0 · 65.0) = 105.4 %mm (1.054 mm)
Wound · MI (E) 82.4 Hz: PUL = 245000000 · 7.50 / (82.4² · 65.0²) · 10 = 640.5
Deliberate divergence: the original's instrument sheets used KF with a fixed coefficient of 150; this calculator applies the monofilament sheet's tiers across the whole page, so the KF column means the same thing wherever you look (see Methodology).
KF uses a tiered coefficient: 150 (d < 100), 156.9767 (100 ≤ d < 150) and 162.6506 (d ≥ 150).
The original read the tension as text with two decimals and used 32-bit Single variables for nylon and gut; this is reproduced as is.
References come from the catalogue row whose value is the largest one less than or equal to the computed value (LOOKUP), with five hypotheses: +0, +2, +3, +4 and +5 %mm (plain) or +0, +15, +30, +45 and +60 PUL (wound).

The original's 5 hypotheses

The original didn't settle for a single reference per string: for every plain series (gut, nylon, KF) and every wound family it offered five catalogue candidates, shifting the calculated value by +0, +2, +3, +4 and +5 %mm for plain strings, or by +0, +15, +30, +45 and +60 PUL for wound ones, looking the catalogue up again at each shift.

Each of the five also yielded the actual tension of that catalogue reference, but the calculation wasn't the same for both string types. For plain strings it came from the plain-string sheet's formula run in reverse: the same formula that gives tension from diameter, applied to the reference's real gauge instead of the theoretical one. For wound strings it came from the wound-string sheet's formula run in reverse (cordes filées), which starts from PUL instead of diameter: the same formula that gives tension from PUL, applied to the reference's real PUL. Either way, that showed how far the real string strayed from the requested tension. This calculator reproduces all five hypotheses of each type, but the results table only shows the first one — the +0 shift, the recommended reference in each cell; the other four have no interface of their own.

What is %mm?What is PUL?Which tension should I use?Methodology History Catalogue