C.M.R. 06, 096, ch. 424, app 096-424-D
1. Introduction
A. XRF performance characteristic sheets and manufacturer's instructions. When an XRF instrument is used for testing paint in residential dwellings and pre-1978 child occupied facilities, it must have a HUD -issued XRF Performance Characteristic Sheet (PCS). XRFs must be used in accordance with the manufacturer's instructions and the PCS. The PCS contains information about XRF readings taken on specific substrates, calibration check tolerances, interpretation of XRF readings, and other aspects of the model's performance.
(1) If discrepancies exist among the PCS, the most stringent guidelines should be followed. For example, if the PCS has a lower (more stringent) calibration check tolerance than the manufacturer's instructions, the PCS should be followed.
2. Inspections in Single-Family Housing. List all testing combinations or each individual surface or component including those that are painted, stained, shellacked, varnished, coated, or wallpaper which covers painted surfaces.
A. Select testing combinations
(1) Perform XRF testing (including the calibration check readings).
(2) Collect and analyze paint-chip samples for testing combinations that cannot be tested with XRF, that have inconclusive XRF results, or for client-approved confirmation of XRF results.
(3) Classify XRF and paint-chip results.
(4) Evaluate the work and results to ensure the quality of the paint inspection.
(5) Document all findings in a report at a minimum meeting the requirements of Section 7(A)(3).
B. Listing testing combinations. Develop a list of all testing combinations in all interior rooms, on all exterior building surfaces, and on surfaces in other exterior areas, and fences, playground equipment, and garages as applicable. An inventory of a house may be completed either before any testing or on a room-by-room basis during testing.
C. Number of room equivalents to inspect. Test all room equivalents inside and outside the dwelling unit. The final report must include a final determination of the presence or absence of lead-based paint on each testing combination in each room equivalent. For varnished, stained, or similar clear-coated floors, measurements in only one room equivalent are permissible if it appears that the floors in the other room equivalents have the same coating.
(1) Some testing combinations have multiple parts. For example, a window testing combination could theoretically be broken down into the interior sill (stool), exterior sill, trough, sash, apron, parting bead, stop bead, casing, and so on. Because it is highly unlikely that all these parts will have different painting histories, usually they should not be considered separate testing combinations unless their professional judgment and field condition dictate otherwise. (Inspectors should regard parts of building components as separate testing combinations if they have evidence that different parts have separate, distinct painting histories). Windows and doors would typically have at least two combinations, interior and exterior.
D. Number of testing combinations to inspect. Inspect each testing combination in each room equivalent, unless similar building component types with identical substrates (such as windows) are all found to contain lead-based paint in the first five interior room equivalents. In that case, testing of that component type in the remaining room equivalents may be discontinued, if and only if the purchaser of the inspection services agrees beforehand to such a discontinuation. The inspector should then conclude that similar building component types in the rest of the dwelling unit also contain lead-based paint. For example, if an inspector finds that baseboards in the first five room equivalents are all positive, the inspector - with the client's permission - may conclude that all remaining room equivalents in the unit contain positive baseboards. This is sometimes referred to as a "positive stop."
(1) Because it is highly unlikely that testing combinations known (and not just presumed) to have been replaced or added to the building after 1977 will contain lead-based paint, they need not be tested. If the age of the testing combination is in doubt, it should be tested.
E. Building component types. Results of an inspection may be summarized by classifying component types across room equivalents if patterns or trends are supported by the data.
F. Substrates. Several types of XRF instruments do not require "substrate correction," needed to correct a systematic bias in an XRF instrument resulting from interference from substrate material beneath the paint. However, all substrates across all room equivalents should be grouped into one of the six substrate categories (brick, concrete, drywall, metal, plaster, or wood) shown on the XRF Performance Characteristic Sheet for the instrument being used. Substrate correction procedures, if required, can then be applied for all building component types with the same substrate. For example, the substrate correction procedure for wooden doors and wooden baseboards can use the same substrate correction value.
3. Number and Location of XRF Readings
A. Number of XRF readings for each testing combination. XRF testing is required for at least one location per testing combination, except for interior and exterior walls, where four readings should be taken, one on each wall. Multiple readings on the same testing combination or testing location are, therefore, unnecessary, except for interior and exterior walls.
(1) Because of the large surface areas and quantities of paint involved, and the possibility of increased spatial variation, take at least four readings (one reading on each wall) in each room equivalent. (For room equivalents with fewer than four walls, test each wall.) For each set of walls with the same painting history in a room equivalent, test the four largest walls. Classify each wall based on its individual XRF reading. If a room equivalent has more than four walls, calculate the average of the readings, round the result to the same number of decimal places as the XRF instrument displays, and classify the remaining walls with the same painting history as the tested walls, based on this rounded average. When the remaining walls in a room equivalent clearly do not have the same painting history as that of the tested walls, test and classify the remaining walls individually. For exterior walls, select at least four sides and average the readings (rounding the result as described above) to obtain a result for any remaining sides. If there are more than four walls and the results of the tested walls do not follow a classification pattern (for example, one is positive and the other three are negative), test each wall individually.
B. Location of XRF readings. The selection of the test location for a specific testing combination should be representative of the paint over the areas that are most likely to be coated with old paint or other lead-based coatings. Thus, locations where the paint appears to be thickest should be selected. Locations where paint has worn away or been scraped off should not be selected. Areas over pipes, electrical surfaces, nails, and other possible interferences should also be avoided if possible. All layers of paint should be included, and the XRF probe faceplate should be able to lie flat against the surface of the test location.
(1) If no acceptable location for XRF testing exists for a given testing combination, a paint-chip sample should be collected and sent to a lead laboratory recognized by NLLAP for analysis of lead in paint. The sample should include all paint layers and should be taken as unobtrusively as possible. Because paint-chip sampling is destructive, a single sample may be collected from a wall and used to characterize the other walls in a room equivalent. For greater reliability, consider collection and analysis of more than one sample.
4. Documentation of XRF Reading Locations. Descriptions of testing combinations must be sufficiently detailed to permit another individual to find them. While it is not necessary to document the exact spot or the exact building component on which the reading was taken, it is necessary to record the exact testing combination measured. Current room uses or colors can change and should not be the only way of identifying them. A numbering system, floor plan, sketch, or other system, is to be used to document which testing combinations were tested.
A. Side identification. Identify perimeter wall sides with letters, numbers, or Roman numerals. For example, Side A for single-family housing is the street side for the address. Side A in multi-family housing is the apartment entry door side. Side B, C, and D are identified clockwise from Side A as one faces the dwelling unit; thus, Wall B is to the left, Wall C is across from Side A, and Side D is to the right of Side A. Each room equivalent's side identification follows the scheme for the whole housing unit. Because a room can have two or more entries, sides should not be allocated based on the entry point. For example, giving a closet a side allocation based on how the room is entered would make it difficult for another person to make an easy identification, especially if the room had two closets and two entryways.
B. Room equivalent identification. Room equivalents should be identified by both a number and a use pattern (for example, Room 5-Kitchen). Room 1 can always be the first room, at the A-D junction at the entryway, or it can be the exterior. Rooms are consecutively numbered clockwise. If multiple closets exist, they are given the side allocation: for example, Room 3, Side C Closet. The exterior is always assigned a separate room equivalent identifier.
C. Sides in a room. Sides in an interior room equivalent follow the overall housing unit side allocation. Therefore, when standing in any four-sided room facing Side C, the room's Side A will always be to the rear, Side B will be to the left, and Side D will be to the right.
D. Building component identification. Individual building components are first identified by their room number and side allocation (for example, the radiator in Room 1, Side B is easily identified). If multiple similar component types are in a room (for example, three windows), they are differentiated from each other by side allocation. If multiple components are on the same wall side, they are differentiated by being numbered left to right when facing the components. For example, three windows on Wall D are identified as windows D1, D2, and D3, left to right. If window D3 has the only old original sash, it is considered a separate testing combination from the other two windows. Codes or abbreviations for building components and/or locations may be used in order to shorten the time needed for data entry. If codes or abbreviations are used, the inspection records and the inspection report must include a table showing their meaning.
(1) A sketch of the dwelling unit's floor plan is required by this protocol. Whatever documentation is used, a description of the room equivalent and testing combination identification system must be included in the final inspection report.
5. XRF Calibration Check Readings. In addition to the manufacturer's recommended warm up and quality control procedures, the XRF operator should take the quality control readings recommended below, unless these are less stringent than the manufacturer's instructions. Quality control for XRF instruments involves readings to check calibration. Most XRFs cannot be calibrated on-site; actual calibration can only be accomplished in the factory.
A. Frequency and number of calibration checks. For each XRF instrument, two sets of XRF calibration check readings are required at least every 4 hours. The first is a set of three nominal-time XRF calibration check readings to be taken before the inspection begins. The second occurs either after the day's inspection work has been completed, or at least every 4 hours, whichever occurs first. To reduce the amount of data that would be lost if the instrument were to go out of calibration between checks, and/or if the manufacturer requires more frequent calibration checks, the calibration check can be repeated more frequently than every 4 hours. If the XRF manufacturer recommends more frequent calibration checks, the manufacturer's instructions should be followed. Calibration should also be checked before the XRF is turned off (for example, to replace a battery or before a lunch break) and after it is turned on again. For example, if an inspection of a large house took 6 hours, there would be three calibration checks: one at the beginning of the inspection, another after 4 hours, and a third at the end of the inspection.
(1) If the XRF is not turned off as the inspector travels from one dwelling unit to the next, calibration checks do not need to be done after each dwelling unit is completed. For example, in multi-family housing, calibration checks do not need to be done after each dwelling unit is inspected, once every 4 hours is usually adequate. Some inspectors do a calibration check between units for two reasons: first, if the instrument goes out of calibration during the inspection of the unit, only that unit needs to be inspected, and, second, if the inspector inadvertently misses a calibration check, the period between checks is less likely to exceed 4 hours.
(2) Some instruments automatically enter a "sleep" or "off" state when not being used continually to prolong battery life. It is not necessary to perform a calibration check before and after each "sleep" state episode unless the manufacturer recommends otherwise.
B. Calibration check standard materials. Portable XRF calibration check readings are taken on the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) or NIST Certified Reference Material using the nominal 1.0 mg/cm2 paint film (or nearly 1.0 in older sets) within the SRM. The complete set of paint films can be obtained by calling (301) 975-2200 or using the NIST SRM site at: http://www.nist.gov/srm/index.cfm.
(1) Calibration checks should be taken through the SRM paint film with the film positioned at least 1 foot (0.3 meters) away from any potential source of lead. The NIST SRM film should not be placed on a toolbox, suitcase, or surface coated with paint, shellac, or any other coating to take calibration check readings. Rather, the NIST SRM film should be attached to a solid (not plywood) wooden board or another non-metal rigid substrate such as drywall or attached directly to the XRF probe. The SRM should be positioned so that readings of it are taken when it is more than 1 foot (0.3 meters) away from a potential source of error. For example, the NIST SRM film can be placed on top of a 1 foot (0.3 meters) thick piece of Styrofoam or other lead-free material, as recommended by the manufacturer before taking readings.
C. Recording and interpreting calibration check readings. Each time calibration check readings are made; three readings should be taken. These readings should be taken using the nominal time which will be used during the inspection, selected from among those specified in the PCS. The open shutter time should be adjusted, if necessary, to reflect the age of the radioactive source. The readings can be recorded on the Calibration Check Test Results form, or stored in the instrument's memory, and printed out or transferred to a computer later. The average of the three calibration check readings should be calculated, rounded to the same number of decimal places as the XRF instrument displays, and recorded on the form.
(1) Large deviations from the NIST SRM value will alert the inspector to problems in the instrument's performance. If the observed calibration check average is outside of the acceptable calibration check tolerance range specified in the instrument's PCS, the manufacturer's instructions should be followed to bring the instrument back into control. A successful calibration check should be obtained before additional XRF testing is conducted. Readings not accompanied by successful calibration checks at the beginning and end of the testing period are unreliable and should be repeated after a successful calibration check has been made. If a backup XRF instrument is used as a replacement, it must successfully pass the initial calibration check test before retesting the affected test locations.
6. Substrate Correction. XRF readings are sometimes subject to systematic biases as a result of interference from substrate material beneath the paint. The magnitude and direction of bias depends on the substrate, the specific XRF instrument being used, and other factors such as temperature and humidity. Results can be biased in either the positive or negative direction and may be quite high.
A. When substrate correction is not required. Some XRF instruments do not need to have their readings corrected for substrate bias on any substrate. Other instruments may only need to apply substrate correction procedures on specific substrates and/or when XRF results are below a specific value. The XRF Performance Characteristic Sheet should be consulted to determine the requirements for a specific instrument and each mode of operation (e.g., nominal time, or time required for intended precision). XRF instruments which do not require correction for any substrate, or require corrections on only a few substrates, have an advantage in that they simplify and shorten the inspection process.
B. Substrate correction procedure. XRF results are corrected for substrate bias by subtracting a correction value determined separately in each house for each type of substrate where lead paint values are in the substrate correction range indicated on the XRF Performance Characteristic Sheet (PCS). In single-family housing, the substrate correction value is determined using the specific instrument(s) used in that house. The correction value (formerly called "Substrate Equivalent Lead" or "SEL") is an average of six XRF readings, with three taken from each of two test locations that have been scraped visually clean of their paint coating. The locations selected for removal of paint should have an initial XRF reading on the painted surface of less than 2.5 mg/cm2, if possible. If all initial readings on a substrate type are greater than 2.5 mg/cm2, the locations with the lowest initial reading should be chosen. Because available data indicate that surfaces with XRF readings in excess of about 3.0 mg/cm2 or 4.0 mg/cm2 are almost always coated with lead-based paint, and since bleed-through of lead into the substrate may occur, or pipes and similarly interfering building components may be behind the material being evaluated, locations with such high readings should be avoided for substrate correction.
(1) After all, XRF testing has been completed but before the final calibration check test has been conducted, XRF results for each substrate type should be reviewed. If any readings fall within the range for substrate correction for a particular substrate, obtain the substrate correction value.
(2) On each selected substrate requiring correction, two different testing combinations must be chosen for paint removal and testing. For example, if the readings are inconclusive for some wooden baseboards, select two baseboards, each from a different room. If some wooden doors also require substrate correction, the inspector should take substrate correction readings on one door and one baseboard. Selecting the precise location of substrate correction should be based on the inspector's ability to remove paint thoroughly from the substrates, the similarity of the substrates, and their accessibility. The XRF probe faceplate must be able to be placed over the scraped area, which should be completely free of paint or other coatings.
(3) The size of the area from which paint is taken depends on the size of the analytical area of the XRF probe faceplate; normally, the area is specified by the manufacturer. To ensure that no paint is included in the bare substrate measurement, the bare area on the substrate should be slightly larger than the analytical area on the XRF probe faceplate.
(4)In all, six readings must be taken for each substrate type that requires correction. All six must be averaged together. Take three readings on the first bare substrate area. Record the substrate and XRF readings on the "Substrate Correction Values" form. Repeat this procedure for the second bare substrate area and record the three readings on the same form. Substrate correction values should be determined using the same instrument used to take readings on the painted surfaces. If more than one XRF model was used to take readings, apply the substrate correction values as specified on each instrument's PCS.
(5) Compute the correction value for each substrate type that requires correction by computing the average of all six readings as shown below and recording the results on the "Substrate Correction Values" form. The formula given below should be used to compute the substrate bias correction value for XRF readings taken on a bare substrate that is not covered with NIST SRM film. A different formula should be used when SRM film must be placed over the bare substrate. The PCS specifies when this correction is necessary and provides the formula for computing the correction value.
(6) Negative values. If more than 20 percent of the corrected values are negative, the instrument's lead paint readings and/or the substrate readings are probably in error. Calibration should be checked, and substrate measurements should be repeated.
(7) Discarding readings. If the manufacturer's instructions call for the deletion of readings at specific times, only readings taken at those specific times should be deleted. Similarly, readings between a successful calibration check and a subsequent unsuccessful calibration check must be discarded. Readings should not be deleted based on any criteria other than what is specified by the manufacturer's instructions.For example, a manufacturer may instruct operators to discard the first XRF reading after a substrate change. If so, only the first reading should be discarded after a substrate change.
7. Classification of XRF Results. XRF results are classified as positive, negative, or inconclusive.
A. A positiveclassification indicates that lead is present on the testing combination at or above the standard of 1.0 mg/cm2. A positive XRF result is any value greater than the upper bound of the inconclusive range, or greater than or equal to the threshold, as specified on the applicable XRF Performance Characteristic Sheet (PCS).
B. A negative classification indicates that lead is not present on the testing combination at or above the standard. A negative XRF result is any value less than the lower bound of the inconclusive range, or less than the threshold, specified on the PCS.
C. An inconclusive classification indicates that the XRF cannot determine with reasonable certainty whether lead is present on the testing combination at or above the standard. An inconclusive XRF result is any value falling within the inconclusive range on the PCS (including the boundary values defining the range). In single-family housing, all inconclusive results should be confirmed by analysis by a laboratory recognized by EPA, under NLLAP, for analysis of lead in paint, unless the client wishes to assume that all inconclusive results are positive.
D. Positive, negative, and inconclusive results apply to the actual testing combination and to any repetitions of the testing combination that were not tested in the room equivalents. Positive results also apply to similar component types in room equivalents that were not tested. For example, suppose that one baseboard in a room equivalent is tested, and that the inspector decided that all four baseboards are a single testing combination. The single XRF result applies to all four baseboards in that room equivalent.
E. When an inconclusive range is specified on the PCS, the inconclusive range includes its upper and lower bounds. XRF results are classified as positive if they are greater than the upper boundary of the inconclusive range, negative if they are less than the lower boundary of the inconclusive range, or inconclusive otherwise.
F. Different XRF models have different inconclusive ranges, depending on the specific XRF model and the mode of operation. The inconclusive range may also be substrate specific. In some cases, the upper and lower limits of the inconclusive range are equal; that value is called the threshold. If the reading is less than the threshold, then the reading is considered negative. If the reading is equal to or greater than the threshold, then the reading is considered positive. Use of the inconclusive range and threshold is detailed in the performance characteristic sheet. The categories include substrate-corrected results, if substrate correction is indicated. XRFs with only threshold values listed on the PCS are advantageous in that classifications of results are either positive or negative (no XRF readings are inconclusive). Note that the final inspection report should not list inconclusive readings as a third category in addition to positive and negative. There are two options for addressing inconclusive readings:
(1) A paint chip may be sampled and sent to a laboratory recognized by EPA, under NLLAP, for analysis of lead in paint.
(2) If the client agrees, all inconclusive readings may be assumed to be positive. It is not permissible to assume any inconclusive reading is negative.
8. Evaluation of the Quality of the Inspection
A. Repeated testing of ten surfaces. Data from HUD's private housing lead-based paint hazard control program show that it is possible to successfully retest painted surfaces without knowing the exact spot which was tested. Select 10 testing combinations at random from the already compiled list in the "Single-Family Housing LBP Testing Data Sheet" for retesting. If possible, the same XRF instrument used in the original inspection should be used in the retesting. If the XRF instrument used in the original inspection is not available and cannot be returned to the site, use an XRF of the same model for retesting. Use the same procedures to retest the 10 testing combinations. The 10 repeat XRF results are compared with the 10 XRF results previously made on the same testing combinations.
(1) The repeat readings and the original readings should not be corrected for substrate bias for the purpose of this comparison. The average of the 10 repeat XRF results should not differ from the 10 original XRF results by more than the retest tolerance limit. The procedure for calculating the retest tolerance limit is specified in the PCS.If the limit is exceeded, the procedure should be repeated using 10 different testing combinations. If the retest tolerance limit is exceeded again, the original inspection is considered deficient.
9. Inspections in Multi-family Housing. For inspection purposes only, multi-family housing is defined as any group of more than four units that are similar in construction from unit to unit.
Selection of housing units, common areas, and exterior site areas. The first step in selecting housing units is to identify buildings in the development with a common construction based on written documentation or visual evidence of construction type. Such buildings can be grouped together for sampling purposes. For example, if two buildings in the development were built at the same time by the same builder and appear to be of similar construction, all of the units in the two buildings can be grouped for sampling purposes, as can the common areas, and exterior site areas. Units can have different sizes, floor plans, and number of bedrooms and still be grouped.
Number of Units to be Tested in Multi-family Building or Developments
Number of Similar Units, Similar Common Areas, or Similar Exterior Sites | Pre-1960 or Unknown-Age Building or Development: Number of Units to Test | 1960-1977 Building or Development: Number of Units to Test |
1-10 | All | All |
11-13 | All | 10 |
14 | All | 10 |
15 | All | 12 |
16-17 | All | 13 |
18 | All | 14 |
19 | All | 15 |
20 | All | 16 |
21-26 | 20 | 16 |
27 | 21 | 17 |
28 | 22 | 18 |
29 | 23 | 18 |
30 | 23 | 19 |
31 | 24 | 19 |
32 | 25 | 19 |
33-34 | 26 | 19 |
35 | 27 | 19 |
36 | 28 | 19 |
37 | 29 | 19 |
38-39 | 30 | 20 |
40-48 | 31 | 21 |
49-50 | 31 | 22 |
51 | 32 | 22 |
52-53 | 33 | 22 |
54 | 34 | 22 |
55-56 | 35 | 22 |
57-58 | 36 | 22 |
59 | 37 | 23 |
60-69 | 38 | 23 |
70-73 | 38 | 24 |
74-75 | 39 | 24 |
76-77 | 40 | 24 |
78-79 | 41 | 24 |
80-88 | 42 | 24 |
89-95 | 42 | 25 |
96-97 | 43 | 25 |
98-99 | 44 | 25 |
100-109 | 45 | 25 |
110-117 | 45 | 26 |
118-119 | 46 | 26 |
120-138 | 47 | 26 |
139-157 | 48 | 26 |
158-159 | 49 | 26 |
160-177 | 49 | 27 |
178-197 | 50 | 27 |
198-218 | 51 | 27 |
219-258 | 52 | 27 |
259-279 | 53 | 27 |
280-299 | 53 | 28 |
300-379 | 54 | 28 |
380-499 | 55 | 28 |
500-776 | 56 | 28 |
777-939 | 57 | 28 |
940-1004 | 57 | 29 |
1005-1022 | 58 | 29 |
1023-1032 | 59 | 29 |
1033-1039 | 59 | 30 |
1500 | 87 | 44 |
2000 | 116 | 58 |
2500 | 145 | 73 |
3000 | 174 | 87 |
3500 | 203 | 102 |
4000 | 232 | 116 |
(1) The specific units to be tested should be chosen randomly from a list of all units in each building or buildings. (For brevity, just "units" are mentioned in describing the random selection procedure, but the procedure is the same for similar units, similar common areas, and similar exterior sites.) The "Selection of Units" form or a comparable form may be used to aid in the selection process. A complete list of all units in each group should be used and a separate identifying sequential number must be assigned to each unit. For example, if apartment addresses are shown as 1A, 1B, 2A, 2B etc., they must be given a sequence number (1, 2, 3, 4, etc.).
Note: Units without identifiers cannot not be selected for inspection and would bias the sampling. The list of units should be complete and verified by consulting building plans or by a physical inspection of the development.
(2) Specific units to be tested should be selected randomly.
Note: Tables of random numbers are often included in statistics books. Today's common full-function computer spreadsheet software products have random number generator functions of sufficient quality for use in lead-based paint inspections.
In Microsoft Excel the code = RAND BETWEEN (X,Y), where X is the first unit number in the list and Y is the final unit number in the list, will generate a random whole number that can be used to randomly select a unit for sampling. The Rand between function is also available in other software, such as Open Office.
Inspectors are, therefore, advised to use them to obtain the random numbers, which can then be used to select the specific numbered units.
(a) The "Selection of Units" form is completed by filling in as many random numbers as are needed in the appropriate column. Numbers for the third column are obtained by multiplying the total development size by each random number. Numbers for the fourth column are obtained by rounding up from the previous calculation to the next whole number. If the whole number in the fourth column has already been selected, that selection should not be entered again. The notation "DUP" should be entered to show that the selection was a duplicate. This process should continue until the required number of distinct sample numbers has been selected. Common areas and exterior room equivalents should be identified at this time, but they are not considered to be separate units.
B. Listing testing combinations and common areas. The "Multi-family Housing LBP Testing Data Sheet" form or a comparable form should be used to list the testing combinations in each unit, common area and exterior site that was selected for inspection. In multi-family housing, the inventory of testing combinations often will be similar for units that have the same number of bedrooms. The inspector should, however, list testing combinations that are unique to each tested unit. For example, some units may contain built-in cabinets while others do not. The selection of testing combinations should, therefore, be carried out independently in each inspected unit.
(1) As in single family housing, take readings on all testing combinations in all room equivalents in each unit selected for testing. However, common areas need to be identified and tested as well.
C. Common areas. Similar common areas and similar exterior sites must always be tested, but in some cases, they can be sampled in much the same way that dwelling units are. Common areas and building exteriors typically have a similar painting history from one building to the next. In multi-family housing, each common area (such as a building lobby, laundry room, or hallway) can be treated like a dwelling unit. If there are multiple similar common areas, they may be grouped for sampling purposes in exactly the same way as regular dwelling units are. However, dwelling units, common areas and exterior sites cannot all be mixed together in a single group.
(1) All testing combinations within each common area or on building exteriors selected for testing must be inspected. This includes playground equipment, benches and miscellaneous testing combinations located throughout the development. The specific common areas and building exteriors to test should be randomly selected, in much the same way as specific units are selected using random numbers.
(2) The number of common areas to test should be taken from the above table. In this instance, common areas and building exteriors can be treated in the same way as housing units (although they are not to be confused with true housing units).
10. Classification of XRF Results in Multi-family Housing. The inspector should record each XRF reading for each testing combination on the "Multi-family Housing LBP Testing Data Sheet," or a comparable form, and indicate whether that testing combination was classified as positive, negative, or inconclusive as described previously for single-family housing.
A. When the inspection is completed in all the selected units and the classification rules have been applied to all XRF results, the HUD "Multi-family Housing: Component Type Report" form or a comparable form should be completed. Building component types groups of like components constructed of the same substrate in the multi-family housing development are aggregated on this form. For example, grouping all interior walls would create an appropriate component type if all walls are plaster. Grouping all doors would not be appropriate; however, if some doors are metal and some are wood. At least 40 testing combinations of a given component type in a multi-family housing development must be tested to obtain the desired level of confidence in the results for that component type. If fewer than 40 testing combinations of a given component type were tested, test additional combinations of that component type. If fewer than 40 components of a given type exist in the units to be tested, test all of the components that do exist.
B. In some cases, additional sampling of the specific component may not be necessary. If no lead at or above the standard is found on that component type, additional measurements should be taken in other units to increase the sample size to 40. However, if all or most of the sampled component types are positive, no further sampling is needed, provided that the building owner agrees with this reduction of testing. For example, if 20 out of 60 doors are tested, and the majority is positive for lead-based paint, all similar doors in the buildings may be presumed positive; only those doors tested and found negative may be treated as negative. Note that the inspector and owner may not presume a component is negative. All required XRF testing and/or laboratory analysis must be completed to conclude that any or all components included in a given component type are negative.
C. The substrate and the component for each component type should be recorded on the HUD "Multi-family Housing: Component Type Report" or comparable form under the heading "Description" (for example, wooden interior doors), as should the total number of testing combinations included in the component type. In addition, for each component type, the aggregated positive, negative, and inconclusive classifications should be recorded as described below. Record the number and percentage of testing combinations classified as:
(1) Positive for lead-based paint. This is based upon a positive XRF reading in accordance with the XRF's Performance Characteristic Sheet;
(2) Low Inconclusive for lead-based paint. This is based on having XRF readings less than the midpoint of the XRF's inconclusive range (if the XRF instrument does not have an inconclusive range (that is, it has a threshold value), this aggregation element should not be provided);
(3) High Inconclusive (high) for lead-based paint. This is based on having XRF readings equal to or greater than the midpoint of the XRF's inconclusive range (if the XRF instrument does not have an inconclusive range (that is, it has a threshold value), this aggregation element should not be provided); and
(4) Negative for lead-based paint.
(5) Positive: Lead-based paint is present on one or more of the components.
(6) Negative: Lead-based paint is not present on the components throughout the development. (Lead may still be present at lower loadings and hazardous leaded dust may be generated during modernization, renovation, repair, remodeling, maintenance, painting or other disturbances of painted surfaces.)
D. The decision that lead-based paint is not present throughout the development is reached if:
(1) 100 percent of the tested component types are negative, or
(2) 100 percent of the tested component types are classified as either negative or inconclusive and all of the inconclusive classifications have XRF readings less than the midpoint of the inconclusive range for the XRF in use.
(3) Note: that the midpoint of the inconclusive range is not a threshold; it is used only for classifying XRF readings in multi-family housing in conjunction with information about other XRF readings as described here.
(4) For cases with greater than or equal to 5% positives and less than 15% positives, as well as no positives but greater than 15% high inconclusive, some confirmatory laboratory testing may be needed to reach a final conclusion, unless the client wishes to assume the validity of the XRF results and that all inconclusive are positive.
(a) For each testing combination with an inconclusive XRF reading at or above the midpoint of the inconclusive range, a paint-chip sample should be analyzed by a laboratory recognized by the EPA NLLAP for the analysis of lead in paint.
(b) If all the laboratory-analyzed samples are negative, it is not necessary to test inconclusive XRF results below the midpoint of the inconclusive range.
(c) If, however, any laboratory results are positive on a component type, all inconclusive equal to or above the midpoint of the inconclusive range should be analyzed, or they should be presumed to be positive.
(d) Percentages of positive or inconclusive results are computed by dividing the number in each classification group by the total number of testing combinations of the component type that were tested. For example, if 245 wooden doors in a multi-family housing development were tested and 69 were classified as inconclusive with XRF readings less than the midpoint of the inconclusive range, 28 percent [(69 / 245) x 100 percent = 28.2 percent] should be recorded on the form in the "<1.0 percent" columns under the heading "Inconclusive."
11. Unsampled Housing Units. If a particular component type in the sampled units is classified as positive, that same component type in the un-sampled units is also classified as positive. For those cases where the number of positive components is small, further analysis may determine if there is a systematic reason for the specific mixture of positive and negative results.
Note: For example, suppose that a few porch railings tested negative, but most tested positive. Examination of the sample results in conjunction with the building records showed that the porch railings classified as positive were all original and the railings classified as negative were all recent replacements. The records did not reveal which units had replaced railings, and due to historic preservation requirements, the replacement railings were identical in appearance to the old railings. Thus, all un-sampled original porch railings could be classified as positive, and all un-sampled recently replaced porch railings could be classified as negative if at least 40 of the replaced porch railings had been tested or were determined to have been installed after 1978.
A. Fewer than 5% positive results. Where a small fraction of XRF readings, less than 5% percent, of a particular component type are positive, several choices are available:
(1) First, the inspector may confirm the results by laboratory analysis, which is considered definitive when performed; a laboratory lead result of 1.0 mg/cm2 or greater (or 0.5 percent by weight or greater) is considered positive.
(2) Second, the inspector may select a second random sample (using un-sampled units only) and test the component type in those units. If less than 2.5% of the combined set of results is positive, the component type may be considered as having lead-based paint in isolated locations, but not having lead-based paint development-wide, with a reasonable degree of confidence. Individual components that are classified positive should be considered as being lead-based painted and managed or abated appropriately.
(3) Finally, if the client chooses not to confirm the results by laboratory analysis and not to take a second set of measurements, then the component type should be considered as having lead-based paint development-wide. The inspector may wish to advise the client that the cost of additional XRF testing or laboratory analysis is usually much less than the cost of lead abatement or interim control projects. This is of particular interest in the situation where few results are positive, because there is a significant chance that the paint, development-wide, may not be lead-based.
B.
C.M.R. 06, 096, ch. 424, app 096-424-D