building.measurement.joint_insulation

Sound reduction index of joints filled with fillers or seals (ISO 10140-1:2021 Annex J).

A joint is a line, not an area: a slit with or without a filler, a foam or sealing tape, a gasket on the rebate of a door or window. Annex J therefore normalizes the sound it lets through to a metre of joint rather than to a square metre of element. With the joint mounted in a highly insulating element, the level difference measured to ISO 10140-2 gives the sound reduction index of joints per metre,

with $S_{\mathrm{n}}=1$ m², $l_{\mathrm{n}}=1$ m, the joint length $l$ and the equivalent absorption area $A$ of the receiving room (joint_sound_reduction_index).

Flanking through the test element. The element the joint sits in is part of every measurement, so J.1 asks for the maximum the arrangement can show, $R_{\mathrm{s},\max }$, measured with the joint sealed on both sides. Unless it lies 10 dB or more above the measured $R_{\mathrm{s}}^{\prime }$, the result is corrected by the rules of ISO 10140-2:2021 A.3 (lab_joint_insulation):

• a difference of 6 dB up to 10 dB: Formula (J.2), $R_{\mathrm{s}}=-10\lg (10^{-R_{\mathrm{s}}^{\prime }/10}-10^{-R_{\mathrm{s},\max }/10})$;
• less than 6 dB: a fixed correction of 1,3 dB, which is what Formula (J.2) gives at 6 dB, and the value is a minimum;
• $R_{\mathrm{s}}^{\prime }$ above $R_{\mathrm{s},\max }-3$ dB: the lower limit of $R_{\mathrm{s}}$ may be set to $R_{\mathrm{s},\max }$ itself, presented in brackets as a minimum value, e.g. $(R_{\mathrm{s}}\ge 50,4$ dB).

Single numbers. $R_{\mathrm{s},\mathrm{w}}$ ($C$; $C_{\mathrm{tr}}$) of ISO 717-1:2020 on the 16 bands 100 Hz to 3 150 Hz, with $C_{100\text{-}5000}$ and $C_{\mathrm{tr},100\text{-}5000}$ when the bands reach 5 000 Hz, as the form of Figure J.7 prints them. Where a band is above $R_{\mathrm{s},\max }-3$ dB, J.1 rates the curve a second time with those indicative bands taken as infinitely high; a difference of more than 1 dB between the two puts the single numbers in brackets too.

The arrangement. J.2.1 asks for a joint longer than 1 m and no wider than 50 mm, J.2.2 for at least 5,0 m of a gap between the parts of a window or door (check_joint_test_element), whose width is read at four positions or more that may not differ by more than 0,3 mm (check_gap_width). A variable slit is measured at its nominal gap width $b_{\mathrm{n}}$ (5 mm when the manufacturer gives none), at the minimum width under a force of normally 100 N/m, and at $b_{\mathrm{n}}+3$ mm (J.4, check_joint_gap_series), and reported as single numbers against the gap width (joint_gap_series, Figures J.8 and J.9).

Citations are to ISO 10140-1:2021 (third edition) Annex J and, for the flanking rules it calls in, to ISO 10140-2:2021 A.3.

Auto-generated from the source docstrings by scripts/generate_api_docs.py (make api-docs). Do not edit by hand.

check_gap_width(readings_mm: ArrayLike) -> GapWidthCheck

Check the gap width read along the joint and average it (J.2.2).

“Determine the gap width at a minimum of four positions, evenly distributed over the total length of sealing. The results shall not deviate by more than 0,3 mm, otherwise readjust the mounting. The average value is denoted as gap width, b.” The deviation is read between the readings, the largest minus the smallest, which is the stricter of the two ways the sentence can be read: a deviation of each reading from the average would allow twice the spread. Whether the positions are evenly distributed is not judged, since only the widths are given.

Parameters

Returns: A GapWidthCheck.

Raises

check_joint_gap_series(series: JointGapSeries) -> JointGapSeriesCheck

Check that a variable slit was measured at the three gap widths of J.4.

“For variable slits, as with openable windows or doors, the test shall be repeated for the following gap width”: a) the nominal gap width $b_{\mathrm{n}}$, b) the minimal gap width $b_{\min }$ at maximal pressure, and c) $b_{\mathrm{n}}+3$, “a gap width 3 mm more than nominal”. Each has to be among the widths of the series.

J.4 gives no tolerance on how close a measured width must come. The gap width of a result is the average of readings that J.2.2 lets spread over 0,3 mm, so a measured width within 0,3 mm of a required one counts as that width. A series that does not name $b_{\min }$ fails J.4 b), because nothing then shows which width it was; pass minimum_gap_mm to joint_gap_series.

Parameters

Returns: A JointGapSeriesCheck.

check_joint_test_element(
    joint_length_m: float,
    joint_width_mm: float,
    *,
    window_or_door_gap: bool = False,
) -> JointTestElementCheck

Check the length and width of a joint before it is tested (J.2.1, J.2.2).

J.2.1: “The length of the joint shall be greater than 1 m and the width of the joint shall be no greater than 50 mm.” J.2.2 adds, for gaps between the parts of windows and doors, that the gap under test “shall have a length of at least 5,0 m with a uniform cross-section”, which may be the sum of several gaps. The 1 m bound is exclusive, the others inclusive.

The uniform cross-section, the shape of Figure J.1 and the design of the environment around the joint (J.2.1 “can only give advice”) carry no number and are not judged here.

Parameters

Returns: A JointTestElementCheck.

Raises

GapWidthCheck(readings_mm: tuple[float, ...])

The gap width read along the joint, and whether the readings agree (J.2.2).

Everything else is read from the readings, so a check cannot be built to pass readings the clause fails.

Attributes

property

Whether there are at least four readings.

property

The average of the readings, the gap width b, in mm.

property

Whether both hold; otherwise J.2.2 says to readjust the mounting.

GapWidthCheck.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot the readings with their average and the 0,3 mm band.

Requires matplotlib (pip install phonometry[plot]); returns the Axes.

property

The largest difference between two readings, in mm.

property

Whether no two readings differ by more than 0,3 mm.

joint_gap_series(
    gap_widths_mm: ArrayLike,
    results: tuple[LabJointInsulationResult, ...] | list[LabJointInsulationResult],
    *,
    nominal_gap_mm: float = 5.0,
    minimum_gap_mm: float | None = None,
) -> JointGapSeries

Collect the results of a variable slit at several gap widths (J.4, J.5).

“For variable slits, as with openable windows or doors, the test shall be repeated” at the nominal gap width $b_{\mathrm{n}}$ given by the manufacturer (“if unknown $b_{\mathrm{n}}$ = 5 mm is to be taken”), at the minimal gap width $b_{\min }$ under the maximal pressure (normally a force of 100 N per metre of sealing) and at $b_{\mathrm{n}}+3$ mm (J.4); NOTE 1 suggests steps of 1 mm from maximum compression until a gap is present. The report gives each of $R_{\mathrm{s},\mathrm{w}}$, $R_{\mathrm{s},\mathrm{w}}+C$ and $R_{\mathrm{s},\mathrm{w}}+C_{\mathrm{tr}}$ as a function of the gap width with the nominal width marked (J.5.2 i)), and the results at $b_{\mathrm{n}}$ and $b_{\mathrm{n}}+3$ are the ones products are compared by (J.5.2 h)). Any set of widths is collected; check_joint_gap_series judges whether the three of J.4 are among them.

Parameters

Returns: A JointGapSeries, sorted by gap width.

Raises

joint_sound_reduction_index(
    l1_db: ArrayLike,
    l2_db: ArrayLike,
    absorption_m2: ArrayLike,
    *,
    joint_length_m: float,
) -> np.ndarray

Sound reduction index of joints per metre, Formula (J.1).

with $S_{\mathrm{n}}=1$ m² and $l_{\mathrm{n}}=1$ m. A joint twice as long lets twice the power through, so the index of a metre is 3 dB above the one a 2 m joint gives without the normalization.

The same formula, with the joint sealed on both sides, gives the maximum of the test arrangement $R_{\mathrm{s},\max }$ that lab_joint_insulation corrects against (J.1).

Parameters

Returns: $R_{\mathrm{s}}$ per band, in dB.

Raises

JointGapSeries(
    gap_widths_mm: np.ndarray,
    results: tuple[LabJointInsulationResult, ...],
    nominal_gap_mm: float,
    minimum_gap_mm: float | None,
    r_s_w_db: np.ndarray,
    r_s_w_c_db: np.ndarray,
    r_s_w_ctr_db: np.ndarray,
)

Single numbers of a variable slit against its gap width (J.4, J.5.2 h), i)).

Attributes

JointGapSeries.at_gap(
    gap_mm: float,
    quantity: Literal['r_s_w', 'r_s_w_c', 'r_s_w_ctr'] = 'r_s_w_ctr',
) -> float

The single number at a gap width, on the lines through the measurements.

Figure J.8 plots “measurement results (open circles) and interpolations (lines plotted through measurement results)”; this reads the straight line between the two widths either side.

Parameters

Returns: The single number, in dB.

Raises

JointGapSeries.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    quantity: Literal['r_s_w', 'r_s_w_c', 'r_s_w_ctr'] = 'r_s_w_ctr',
    **kwargs: Any,
) -> Axes

Plot a single number against the gap width, as Figure J.8.

The measured widths as open circles on the line through them, with $b_{\min }$, $b_{\mathrm{n}}$ and $b_{\mathrm{n}}+3$ marked. Requires matplotlib (pip install phonometry[plot]); returns the Axes.

JointGapSeries.plot_octave_bands(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot $R_{\mathrm{s}}$ in octave bands at every gap width, as Figure J.9.

Every result must hold whole octaves of one-third-octave bands. Requires matplotlib (pip install phonometry[plot]); returns the Axes.

JointGapSeries.single_number(
    quantity: Literal['r_s_w', 'r_s_w_c', 'r_s_w_ctr'] = 'r_s_w_ctr',
) -> np.ndarray

The single number at each width: "r_s_w", "r_s_w_c" or "r_s_w_ctr".

property

(bn, bn + 3): the working range of Figure J.8, in mm.

The figure draws it as $\Delta b$ and its key names it $\Delta b_{\mathrm{n}}$ (recorded in the errata registry); the plot uses the symbol of the drawing. The 3 mm is the one J.4 c) measures at; where J.5.2 h) lets “a gap range other than 3 mm” apply, at_gap reads the single numbers at any other measured width.

JointGapSeriesCheck(
    gap_widths_mm: tuple[float, ...],
    nominal_gap_mm: float,
    minimum_gap_mm: float | None,
)

Whether a variable slit was measured at the three gap widths of J.4.

The three widths J.4 asks for are read from the nominal and the minimal gap width, so the verdicts are not fields: a check cannot be built to pass a series that misses one.

Attributes

property

Whether a measured width is $b_{\min }$ (J.4 b)).

False when the series names no $b_{\min }$, since nothing then shows it was measured.

property

Whether a measured width is $b_{\mathrm{n}}$ (J.4 a)).

property

Whether all three were measured.

JointGapSeriesCheck.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot the measured gap widths against the three widths J.4 asks for.

Requires matplotlib (pip install phonometry[plot]); returns the Axes.

property

How far a measured width may lie from a required one, in mm (0,3 mm, J.2.2).

property

Whether a measured width is $b_{\mathrm{n}}+3$ mm (J.4 c)).

JointOpenBandRating(
    open_frequencies_hz: tuple[float, ...],
    r_s_w_db: int | None,
    c_db: int | None,
    ctr_db: int | None,
    c_100_5000_db: int | None,
    ctr_100_5000_db: int | None,
)

The single numbers with the indicative bands taken as infinitely high (J.1).

Where $R_{\mathrm{s}}^{\prime }$ is above $R_{\mathrm{s},\max }-3$ dB the band says only that the joint is better than the arrangement can show. J.1 rates the curve a second time with an infinitely high sound reduction index in those bands: they then add no unfavourable deviation and no transmitted energy. A term is None when it is unbounded, which happens when every band it sums is indicative, or when the bands do not reach its range.

Attributes

JointOpenBandRating.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot the open-band single numbers beside the bands taken as open.

Requires matplotlib (pip install phonometry[plot]); returns the Axes.

JointTestElementCheck(
    joint_length_m: float,
    joint_width_mm: float,
    window_or_door_gap: bool,
)

Whether a joint is long and narrow enough to be tested (J.2.1, J.2.2).

The bounds are the clause’s, so the verdicts are read from the length and the width and are not fields: a check cannot be built to pass a joint the clause fails.

Attributes

property

Whether the length is greater than 1 m and, for a window or door gap, at least 5,0 m.

The 1 m bound is exclusive and the 5,0 m bound inclusive, read with a relative slack so that a length on it is not failed by rounding.

property

Whether both the length and the width hold.

JointTestElementCheck.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot the length and width of the joint against their bounds.

Requires matplotlib (pip install phonometry[plot]); returns the Axes.

property

Whether the width is no greater than 50 mm.

lab_joint_insulation(
    r_s_measured_db: ArrayLike,
    r_s_max_db: ArrayLike,
    frequencies_hz: ArrayLike,
    *,
    joint_length_m: float | None = None,
    limit_at_maximum: bool = True,
) -> LabJointInsulationResult

Sound reduction index of a joint, corrected for the test arrangement (J.1).

$R_{\mathrm{s}}^{\prime }$ and $R_{\mathrm{s},\max }$ are both Formula (J.1) (joint_sound_reduction_index): the first with the joint as it is tested, the second with it sealed on both sides, “e.g. with elastic sealant”. Band by band, with the difference $d=R_{\mathrm{s},\max }-R_{\mathrm{s}}^{\prime }$:

• $d\ge 10$ dB: no correction, $R_{\mathrm{s}}=R_{\mathrm{s}}^{\prime }$;
• $6\le d<10$ dB: Formula (J.2), $R_{\mathrm{s}}=-10\lg (10^{-R_{\mathrm{s}}^{\prime }/10}-10^{-R_{\mathrm{s},\max }/10})$;
• $d<6$ dB: $R_{\mathrm{s}}=R_{\mathrm{s}}^{\prime }+1,3$ dB, the correction Formula (J.2) gives at 6 dB, a minimum value (ISO 10140-2:2021 A.3);
• $d<3$ dB, with limit_at_maximum (the default): the lower limit set to $R_{\mathrm{s}}=R_{\mathrm{s},\max }$, which J.1 allows (“may”) and presents in brackets as a minimum value.

The last rule is exact rather than conservative: the arrangement passes ${\tau }_{\mathrm{s},\max }$ alone and ${\tau }_{\mathrm{s}}+{\tau }_{\mathrm{s},\max }$ with the joint, so a measured index within 3 dB of the maximum means ${\tau }_{\mathrm{s}}<{\tau }_{\mathrm{s},\max }$.

With the 16 bands 100 Hz to 3 150 Hz the result carries $R_{\mathrm{s},\mathrm{w}}$ ($C$; $C_{\mathrm{tr}}$) of ISO 717-1, and with the 18 bands to 5 000 Hz $C_{100\text{-}5000}$ and $C_{\mathrm{tr},100\text{-}5000}$, the evaluation of Figure J.7. Bands with $d<3$ dB are indicative: J.1 rates the curve again with them taken as infinitely high, and a change of more than 1 dB puts the single numbers in brackets (LabJointInsulationResult.bracketed). The indicative bands are read on $R_{\mathrm{s}}^{\prime }$, the same condition as the bracketed lower limit, whichever rule corrected them.

Parameters

Returns: A LabJointInsulationResult.

Raises

LabJointInsulationResult(
    frequencies_hz: np.ndarray,
    r_s_measured_db: np.ndarray,
    r_s_max_db: np.ndarray,
    r_s_db: np.ndarray,
    joint_length_m: float | None,
    rating: WeightedRatingResult | None,
    c_100_5000_db: int | None,
    ctr_100_5000_db: int | None,
    max_rating: WeightedRatingResult | None,
    open_band_rating: JointOpenBandRating | None,
    *,
    limit_at_maximum: bool = True,
)

Sound reduction index of a joint per metre (ISO 10140-1:2021 Annex J).

Attributes

How each band was corrected (regime) is read from the measured and the maximum indices and limit_at_maximum, so it is not a field.

property

Whether the single numbers are presented in brackets (J.1).

They are when rating the indicative bands as infinitely high moves any of $R_{\mathrm{s},\mathrm{w}}$, $R_{\mathrm{s},\mathrm{w}}+C$, $R_{\mathrm{s},\mathrm{w}}+C_{\mathrm{tr}}$ or their enlarged-range sums by more than 1 dB, or leaves one of them unbounded.

property

Bands where $R_{\mathrm{s}}^{\prime }>R_{\mathrm{s},\max }-3$ dB (J.1).

property

Bands whose $R_{\mathrm{s}}$ is a minimum value ("limit" or "maximum").

LabJointInsulationResult.octave_bands() -> tuple[np.ndarray, np.ndarray]

(octave centres in Hz, Rs,oct in dB) from the three thirds of each octave.

$R_{\mathrm{oct}}=-10\lg [(1/3)\sum 10^{-R_j/10}]$: the transmitted power of the three thirds averaged, as Figure J.9 plots a joint at several gap widths in octave bands.

LabJointInsulationResult.plot(
    ax: Axes | None = None,
    *,
    language: str = 'en',
    **kwargs: Any,
) -> Axes

Plot $R_{\mathrm{s}}$ per band with the maximum of the arrangement.

The test curve with its minimum values marked, $R_{\mathrm{s}}^{\prime }$ and $R_{\mathrm{s},\max }$, and the shifted ISO 717-1 reference curve over the rating range, as the diagram of Figure J.7. Requires matplotlib (pip install phonometry[plot]); returns the Axes.

property

$R_{\mathrm{s},\mathrm{w}}+C$, in dB, or None without a rating.

property

$R_{\mathrm{s},\mathrm{Atr}}=R_{\mathrm{s},\mathrm{w}}+C_{\mathrm{tr}}$, in dB, or None.

property

$R_{\mathrm{s},\mathrm{w}}$, in dB, or None without a rating.

property

How each band is corrected for the flanking of the arrangement (J.1).

With $d=R_{\mathrm{s},\max }-R_{\mathrm{s}}^{\prime }$: "uncorrected" when $R_{\mathrm{s},\max }$ is at least 10 dB above, "corrected" by Formula (J.2) from 6 dB to 10 dB, "limit" (the fixed 1,3 dB) below 6 dB, and "maximum" (set to $R_{\mathrm{s},\max }$) above $R_{\mathrm{s},\max }-3$ dB when limit_at_maximum.

LabJointInsulationResult.report(
    path: str,
    *,
    metadata: ReportMetadata | None = None,
    engine: str = 'reportlab',
    verbose: bool = False,
    language: str = 'en',
) -> str

Render the form of Figure J.7 to a one-page PDF.

The header of the form (client, specimen, date, test length, separation wall, test noise, room volumes, maximum joint sound reduction index, mounting and climate), the table of $R_{\mathrm{s}}$ from 100 Hz to 5 000 Hz beside its diagram, and the evaluation according to ISO 717-1: $R_{\mathrm{s},\mathrm{w}}$ ($C$; $C_{\mathrm{tr}}$), $C_{100\text{-}5000}$ and $C_{\mathrm{tr},100\text{-}5000}$, in brackets when J.1 puts them there. Minimum values are printed with ≥, those set to $R_{\mathrm{s},\max }$ in brackets.

Parameters

Returns: The written path as a str.

Raises