Don't blame Wheelcalc for what is almost certainly an example of GIGO.

Spoke length calculators don't have minds, moods or whims. They NEVER deviate form being simple calculators running a simple formula. All spoke calculators use the same basic formula, with some variation among the various programs to account for the small fudges such as the allowance for the extra length needed for the over/under lacing at the crosses, or for the 1mm difference between the hub's on-center spoke circle measurement vs. the inside elbow measurement of spokes.

So while various calculators will give slightly different results the same calculator is 100% bulletproof reliable IF proper and consistent data is entered. This is why most wheelbuilders will always use the same calculator. Once they build a wheel or two, they know which way to round, or otherwise tweak the calculated length to get the results they prefer every time.

So when a calculator gives a wrong answer to someone used to it, then the only source to the error if wrong data entered. This can be as simple as calculating 2x and building 3x, or most commonly the result of using published specs. for the rim's ERD instead of an actual measurement. Measured or from a spec, remember that the calculator's answer will have the spokes end at the ERD entered, so it's critical to allow for the nipple when entering the data, or add 2-3mm to the result to correct if the rim's actual diameter at the spoke hole was used.

BTW- spoke calculations are also prone to error from typo's, of forgetting to change the flange diameter when switching front to rear on pairs, so check your work before simply accepting the results and lacing a wheel.