Fix five conversion factors that disagree with their definitions

[tmilnthorp]

Five units convert with a factor that doesn't match the unit's definition. This fixes them and updates the test values to the exact reciprocals.

Unit	Old	New	Relative change
Power.BoilerHorsepower	9812.5 W	9809.5 W	−3.1e-4
Energy.ThermEc	1.05505585262e8 J	1.05506e8 J	+3.9e-6
LeakRate.TorrLiterPerSecond	1/7.5 = 0.133333 Pa·m³/s	101325/760 × 1e-3 = 0.1333224 Pa·m³/s	−8.2e-5
Temperature.SolarTemperature	5778 K	5772 K	−1.0e-3
Mass.SolarMass	1.98947e30 kg	1.98847e30 kg	−5.0e-4

Sources and history

BoilerHorsepower. NIST SP 811, Appendix B.8: horsepower (boiler) = 9.80950 E+03 W. The old value came from the original horsepower units in 2015 (#65, "as defined on Wikipedia"); the value itself was never discussed.

There are two boiler horsepower variants (Wikipedia):

• the 1884 ASME definition, about 33,469.5 BTU/h (9809.5 W), which is the NIST value used here;
• current industrial practice, 33,475 BTU/h (about 9811 W).

The old 9812.5 W matches neither. The unit's doc now says which variant it is.

ThermEc. NIST SP 811, Appendix B.8: therm (EC) = 1.05506 E+08 J. The EC therm is defined as 100,000 BTU_ISO = exactly 105,506,000 J (Wikipedia, citing the EC Official Journal L 073 of 1972). The old value was 100,000 BTU_IT instead. ThermUs (1.054804e8) and ThermImperial already match their definitions and are unchanged.

TorrLiterPerSecond. 1 Torr is exactly 101325/760 Pa (NIST SP 811: 1.333224 E+02 Pa), and 1 L = 1e-3 m³. This is the same torr that Pressure.Torr already uses ({x} * 101325 / 760), so today a leak rate in Torr·L/s and a pressure in Torr times a volume flow in L/s disagree by 82 ppm.

The value 7.5 was added with LeakRate in #1222, and verified at the time against sensorsone.com, which shows the rounded figure. Leak rate references commonly quote 1 mbar·L/s ≈ 0.75 Torr·L/s for the same reason, but the exact ratio is 0.750062.

SolarTemperature. IAU 2015 Resolution B3 adopts a nominal solar effective temperature of 5772 K as a conversion constant (Prša et al. 2016, arXiv:1510.07674). The same resolution defines the nominal solar radius and luminosity:

• SolarRadius already moved to the B3 nominal value in Use the Nominal Solar Radius as defined in IAU Res. B3 #1424.
• SolarLuminosity uses the B3 nominal 3.828e26 W.

5778 K is an older literature value. It was added in #683 without a source.

SolarMass. The old value, 1.98947e30, doesn't match any published value. It looks like a typo of 1.98847e30. The unit links to Wikipedia, which currently gives two values:

• Best estimate: (1.988475 ± 0.000092)e30 kg.
• Nominal solar mass: 1.988416e30 kg.

The IAU defines only a nominal mass parameter (GM), so the nominal mass in kg is that parameter divided by the measured G, and it changes when G is revised. This PR uses 1.98847e30, within the uncertainty of the best estimate, as the smallest correction. If you'd prefer the IAU-derived nominal value for consistency with SolarRadius, I'm happy to change it. The value was added in #680.

Docs

Each changed unit's XmlDocSummary now states the value used and its source, and XmlDocRemarks links to that source:

• NIST SP 811 Appendix B.8 for BoilerHorsepower, ThermEc and TorrLiterPerSecond;
• arXiv:1510.07674 for SolarTemperature;
• the existing Wikipedia link for SolarMass.

ThermEc had the generic therm description shared by all three therms; it now describes the EC therm specifically. TorrLiterPerSecond and SolarTemperature had no docs.

Also

Mass.json had a trailing comma after its last unit, which isn't valid JSON. CodeGen's parser tolerates it, but other tools don't.

Not included

A broader audit of the unit definitions found other candidates. They need a decision about which variant to use, so they're left for separate PRs:

• Pressure.MeterOfHead uses a water density of 999.744 kg/m³; MeterOfWaterColumn uses 1000.
• The thermochemical calorie is mixed with the IT BTU in several heat-transfer quantities.
• Speed.Mach uses a fixed 340.29 m/s with no reference conditions.

Testing

• The test values (...InOne<BaseUnit>) are updated to the exact reciprocals of the new factors.
• UnitsNet.Tests passes on net10.0. The only failures are two FeetInchesTests de-CH cases, which fail on master on my machine too and pass on CI.

The regenerated code is a separate commit.

🤖 Generated with Claude Code

https://claude.ai/code/session_012XKhDsyHDc5BHrmibxScqG

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98%. Comparing base (3bf1ef4) to head (00be8c8).

Additional details and impacted files

@@          Coverage Diff           @@
##           master   #1743   +/-   ##
======================================
  Coverage      98%     98%           
======================================
  Files         515     515           
  Lines       24092   24092           
======================================
  Hits        23692   23692           
  Misses        400     400           

Flag	Coverage Δ
net48	97% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.