[69 FR 1785]

Light Scattering – An emitted light beam passes through a defined sample volume.  PM causes incident light scattering which is focused by a detector and converted into an electrical output.  A second detector receives a portion of the scattered light and makes an adjustment for change in emitted power.  Used for continuous sampling.

Beta Gauge - PM mass density is measured using the technique of Beta attenuation.  A small Beta source is coupled to a sensitive detector which counts the Beta particles.  As the mass of particles increases the Beta count is reduced.  The relationship between the decrease in count and the particulate mass is computed according to a known equation.  Used when batch sampling is required.

Acoustic Energy - Shock waves caused by the impact of particles with a probe inserted into the flow are used to measure the particulate concentration.  The number and energy of each impact coupled with knowledge of the flow velocity, allows calculation of the particulate mass.  Used for continuous sampling.

	The PM CEMS must pass 7-day zero and upscale drift checks (No greater than 2%, zero reference is between 0-20% and upscale reference is between 50-100% of the PM CEMS response range).
	If using sample volumes, a sample volume check must be completed and passed daily.
	Any auxiliary monitors needed to convert data must meet calibration drift (No greater than 2%).
	If flue gas flow rates vary by more than 10%, the PM CEMS sample must be isokinetic.

	Minimum of 15 and maximum of 20 manual reference methods are conducted over a full range of PM CEMS responses.
	Additional testing can be required if a source meets criteria outlined in Section 8.8.
	Reference method test runs can be less than 60 minutes in duration (e.g. 20 minutes).
	3 runs or 20% of the total runs must be at 3 PM concentration levels (0-50%, 25-75%, or 50-100% of the maximum flue gas PM concentration).
	Sampling location must meet USEPA Method 1 criteria.
	Reference method sampling will follow USEPA Methods 5, 5I, or 17.
	The correlation model used is dependant on performance criteria for each model outlined in Section 13.2.
	For batch sampling, the response time (or cycle time) must be no longer than 15 minutes.  The delay between the end of sampling and sample analysis reporting must be no greater than 3 minutes.  Sampling time must be no less than 30% of the cycle time.

	Zero and upscale drift checks (Adjust when drift exceeds 4% for 5 consecutive days or 8% for 1 day).
	Sample volume checks (Adjust when error exceeds 10% for 5 consecutive days or 20% for 1 day).

	Challenge the PM CEMS 3 times with each of a 0-20, 40-60, and 70-100 audit standard (Use a separate standard for audit points 1, 2, and 3).
	In compliance if below +/-10% of the average audit value or 7.5% of the standard.

	Independently measure the volume of each batch cycle or time period with a calibrated device on a consistent temperature, pressure, and moisture basis.
	3 measurements for batch extractive monitors.
	Three, 20-minute periods for continuous extractive monitors.
	In compliance if within +/- 5% of the average sample volume audit value.

	Conducted in frequencies specified in the applicable permit.
	May be conducted in place of a RRA.
	Conduct an SVA prior to testing.
	Conduct testing the same as the correlation test except using only 12 test runs.

	Conducted in frequencies specified in the applicable permit.
	Conduct an SVA prior to testing.
	Collect 3 simultaneous reference method PM and PM CEMS measurements.