SOP - Light Scattering Coefficient - Integrating Nephelometer - TSI 3563
From TROPOS Wiki
Return to Standard Operating Procedures
The integrating nephelometer measures the scattering component of the aerosol light extinction coefficient. This variable, also called the scattering coefficient or the volumetric scattering cross-section, is directly related to the climate forcing and visibility reduction caused by aerosols, and consequently is one of the core aerosol variables recommended for GAW stations. The scattering coefficient has dimensions of cross-sectional area per unit volume, i.e., m2 m-3, which is often reported as inverse meters (m-1) or inverse megameters (Mm-1, 1 Mm-1 = 10-6 m-1.)
Most integrating nephelometers, including the TSI model 3563, operate by illuminating a volume of air with a diffuse light source. A photomultiplier tube views a conical section, with its axis perpendicular to the light source, of the illuminated volume. This special geometry provides the angular integration of the scattered light needed to provide a signal that is proportional to the scattering coefficient. The instrument is calibrated by filling the sample volume with gases of known scattering cross-section, commonly CO2 and air. Internal temperature, pressure, and relative humidity sensors measure the environmental state of the sample. More information on the theory of operation of the instrument is given in Chapter 7 of the instrument operating manual. Theory and applications of the instrument are described further in Heintzenberg and Charlson (1996).
The TSI model 3563 measures the scattering coefficient at three wavelengths (450, 550, and 700 nm) and for two ranges of angular integration (“total” scatter 7-170º, “back” scatter 90-170º). The operating characteristics of the instrument, including the wavelength and angular response, were characterized in detail by Anderson et al. (1996). The total scattering coefficient is frequently written as σsp or Bsp, although Bscat is sometimes used. Likewise, the hemispheric backscattering coefficient is often written as σbsp.
This document describes specific procedures that apply to the TSI model 3563 integrating nephelometer. The specific procedures are also applicable to other versions of the TSI nephelometer, which may operate at a single wavelength and do not have a backscatter shutter. Many of the general operating procedures can also be applied to integrating nephelometers from other manufacturers, but those instruments may differ in the specific operating procedures.
This manual was written specifically for users of TSI Model 3563 integrating nephelometers. Its purpose is to ensure the quality of nephelometer measurements by advising users of ways to optimize nephelometer performance, to recognize instrument problems, and to perform simple maintenance and repair procedures. This is not meant to be a comprehensive document in that all potential instrument problems are not addressed here. Most of the common preventative maintenance procedures are discussed in detail in the Instruction Manual that comes with Model 3550/3560 Series Integrating Nephelometers. The most commonly encountered problems and maintenance procedures, however, are discussed here. This manual was written for field technicians of the World Meteorological Organization (WMO) Global Atmosphere Watch (GAW) program, so that a field reference document for instrument maintenance, repairs, and performance checks at remote field sites would be available. The hope is that comparable care for the instruments at different sites will lead to a similar high quality of nephelometer performance and reduced instrument down time for unscheduled maintenance and repairs.
Sampling techniques for the TSI model 3563 are described in Chapter 2 of GAW Report 153. Basically, a sample from undisturbed ambient air must be brought to the instrument with minimal losses of particles in the size range that typically dominates aerosol light scattering (0.1-10 μm diameter). The instrument response is independent of flow rate, and flow rates of 10-30 lpm are typically used to give an adequate time response (1/e-response times of 5-15 seconds). Heating from the lamp causes a warming of 4-5 ºC, but additional heating, drying, or insulation may be needed in humid environments, particularly in air conditioned laboratories, to achieve the desired sample relative humidity of 30-40%.
The nephelometer should be configured using the following commands (these are described in Chapter 6 of the Nephelometer Instruction Manual):
STB61 (sufficient for a high flow rate like 30 lpm, should be longer for lower flow rates)
In this configuration, the nephelometer measures the scattering coefficient of filtered air for 54 minutes of each hour and reports 1-minute average values. There is a 5 minute zero period, and two 61-second blanking periods at the start and end of each zero check.
Under extremely clean conditions, such as at the South Pole, with aerosol scattering coefficients frequently below 1 Mm-1, it may be desirable to average the results of multiple zero checks. This can be accomplished by issuing the "SMZ" command with a larger value than 1, e.g., "SMZ4" would cause the nephelometer to use the average the last four zero checks.
The time resolution of nephelometer measurements should allow calculation of hourly statistics (mean, standard deviation, selected percentiles), which implies that the data be logged as averages over 1-3 minutes. The primary parameters needed for using and interpreting data from the TSI integrating nephelometer include all six scattering channels, temperature, pressure, and relative humidity. Supporting data should be recorded regularly (at least hourly) to allow data quality assessments, including the instrument status flags, lamp voltage and current, background signals from zero-air checks, and reference calibrator count rates.
In order to track the performance of a nephelometer, records should be kept of diagnostic measurements over time. This is the best way to determine if the performance of the nephelometer has changed. Span gas and overnight noise checks (discussed below) document the stability of instrument calibration and performance over time. Zero background checks show when the instrument background changes, and are especially useful in showing when the inside of a nephelometer is getting dirty. The monitoring of lamp current and voltage is necessary because lamps draw more current as they age. If the lamp draws too much current, the analog circuit board could be damaged. We recommend replacing the lamp when the lamp current rises above about 6.5 amps, as experience has shown that lamp failure is likely within about a week. Temperature, pressure and relative humidity measurements are required for interpretation of nephelometer measurements, and are also useful in diagnosing many potential instrument problems.
Measurements and checks that should be recorded and monitored over time include:
- Daily. Review measurements of scattering coefficient, instrument temperature, pressure, relative humidity, lamp voltage and lamp current for indications of abnormal operation (spikes, level jumps, drift). Review hourly zero background checks for stability. Abnormal operation should be noted in the instrument log and may require data flagging, editing, or removal prior to archival; such notes should be filed with other instrument metadata.
- Weekly to Monthly. Span gas checks and leak checks should be conducted at least monthly, and more frequently if excess variability is noted in the monthly results.
- Annually. Overnight noise checks should be performed at least annually.
The nephelometer is calibrated by measuring the scattering by two different gases with know scattering coefficients, typically air (low) and CO2 (high). Detailed instructions on how to calibrate the nephelometer are given in the TSI operating manual. Normally, the instrument calibration is checked by measuring the scattering by air and CO2 as if they were sample air, following the "span check" procedure in the Appendix. A full calibration, including adjustment of the internal calibration constants, should be performed only when a span check or instrument comparison suggests that a nephelometer’s calibration has shifted. Routine re-calibration is not recommended as long as regular span checks are performed. The TSI nephelometer software displays the K2 and K4 constants determined in each calibration. The K2 constant is a measure of how much light is being detected by each PMT during the calibration portion of each chopper cycle. This value can vary over a fairly wide range depending on the thickness or on the presence of scratches in the finish of the reflective coating on the chopper shutter. Typical values for K2 for all three wavelengths in a properly functioning nephelometer are 2E-3 to 8E-3, although it is possible that values for a particular nephelometer could lie slightly outside this range. The K4 constant is related to the fraction of the scattering volume illuminated during the backscatter measurement. Typically, the value of this constant is near 0.5.
After a calibration has been performed, it is always a good idea to perform a span gas check to verify that the calibration results are reproducible. If the span check errors are large, a repeat of the calibration may be necessary. Alternatively, the full calibration procedure can be repeated until reproducible values of the K2 and K4 constants are achieved.
The GAW World Data Center for Aerosol Physics (http://gaw.tropos.de/WCCAP/index.html) includes a review of nephelometer operation during station audits. The audit checks the inlet system, instrument logbooks, physical instrument condition, and calibration (or span check) procedures. If time permits, the instrument noise level is determined by operating the instrument with a filter on the inlet for several hours.
Light scattering, instrument and environmental parameter data from the nephelometer may be retrieved continuously, via the communication port, using an external data system (e.g. a station’s central data system) or the data logging application provided by the manufacturer.
In either case the raw data files should be retained and archived at the station or central facility. Data backup and processing at the station or central office should be carried out regularly, with a recommended minimum frequency of weekly.
Minimum requirements for data examination include plotting the time series of light scattering data and examination of the station’s nephelometer log to identify possible sample contamination or instrumental problems, such as zero drift. Any such episodes should be flagged in the subsequent edited data file and excluded from processing in the hourly statistics. Other levels of processing may be applied, for example zero and span drift correction if appropriate. If these procedures are carried out, this information should be added to the metadata file for this period and form part of the metadata submission to the GAW archive. At some stations/locations data are screened using selected wind-sectors or species concentration data e.g. radon, to ensure that measurements represent specifically defined air masses. A description of the selection procedures should also form part of the submitted metadata file. (Some limited information on screening procedures is given in the GAW Aerosol Procedures manual, GAW 153, p 6, 2003, and more detailed information can be obtained from the Aerosol SAG). Higher level processing, for example to include truncation errors may also be carried out using the relevant instrument transfer function when this is known, any such processing should be indicated in the metadata file.
The GAW aerosol sampling manual recommends minimum sampling rates of 1 to 3 minutes and the minimum statistics for light scattering data are arithmetic mean and standard deviation, median, 5 and 95 percentiles for each hour (GAW 153, p 28). Light scattering data should be adjusted to standard temperature and pressure (0ºC, 1013.25 hPa) prior to submission. The hourly data statistics should also include a flag to indicate corrupt or contaminated data.
Edited hourly statistics files, including all relevant metadata should be prepared into the format required by the GAW archive and submitted regularly to the GAW aerosol data center. These data should also be retained and archived at the station or central facility.
Maintenance and Repair
New Instrument Checkout
Initial Inspection and Tests
A Model 3563 nephelometer arriving new from the TSI factory will most likely be in excellent condition. TSI ships these nephelometers in large wooden crates which are form-fitted with blown-in foam. Even with careful packing, however, some instrument components can loosen if the crate is handled roughly. Upon receipt of a new instrument, the following items should be inspected. This inspection will require the removal of the nephelometer photomultiplier and top covers.
- Photomultiplier Tubes (PMTs). With the power cord disconnected, open the PMT housing by removing the four PMT cover screws and remove the PMT cover. Reseat each PMT, wiggling the tube to ensure a good firm fit into the socket. TSI now puts a dab of silicone adhesive at the base of the PMT housings to fix them to the optical block base and prevent them from falling out during shipment. Older nephelometers did not have this adhesive applied. Even with the dab of adhesive on the PMT housings, however, they can still move sideways so that the light path might not be fully centered on the PMT window. Afterwards, be sure to replace the PMT housing before connecting the power cord to the nephelometer. Applying power to the photomultiplier tubes with the PMT housing removed may permanently damage the PMTs.
- Electrical and tubing connections. It is unlikely that any of the electrical or tubing connections would come loose during shipping, but it is a good idea to check them anyway. All of the cable-to-cable and cable-to-circuit board electrical connectors should be checked to make sure they are not loose. In addition, check for a firm connection on the electrical connector that joins the two circuit boards. The 1/2-inch nylon Swagelok nuts between the HEPA zero air filter and the instrument body should be checked for tightness and the 1/4-inch silicone tubing at the vent ports should also be checked to ensure that it is securely connected.
Model 3563 nephelometers are calibrated just before they leave the factory so it is not recommended to recalibrate the instrument unless a performance check suggests a problem with either the instrument or the calibration. There are two performance checks that can be done upon receipt of a new nephelometer. The first is a span gas check, described in detail in the Appendix (6.1).
In a span gas check, the scattering coefficients of a low span gas (typically filtered air) and a high span gas (for example, CO2) are measured under instrument conditions of temperature and pressure. The results are used to derive the measured scattering coefficient of CO2 under conditions of standard temperature and pressure (STP; 273.15K and 1013.25 mb). The measured value of scattering by pure CO2 is compared with the published value [Anderson et al., 1996; Anderson and Ogren, 1998] for each measurement wavelength. The mean error in the CO2 measurement (i.e., the difference from the CO2 target value), calculated from each of the six nephelometer channels (three wavelengths each with a total and hemispheric backscatter measurement) should be within a few percent, with no individual channel's error being larger than 10%. If observed errors are larger than this, it suggests an instrument problem and/or a poor calibration. A span check algorithm is provided in Appendix A so that users can perform these calculations. As discussed below, span gas checks should occur at regular intervals (e.g., weekly to monthly) so that instrument performance can be tracked over time.
Span checks that show large negative values are often caused by CO2 either not entering the nephelometer as expected or not staying inside the instrument. If the CO2 is delivered under elevated pressure, hoses can be blown off fittings inside the nephelometer cover. Check to make sure no tubes have been disconnected or ruptured and that CO2 is in fact flowing through the nephelometer. Since the CO2 measurement is made relative to the measurement of filtered air, large negative errors will also be encountered if the filtered air measurement is compromised. This can happen if the zero filter ball valve is not completely sealing off the inlet and directing all air through the HEPA filter. If this turns out to be the case, either adjust the ball valve so that it completely seals off the inlet, or else replace it if necessary.
The second performance check is an instrument noise check. For this check, a second HEPA filter is required and should be mounted on the instrument inlet. Nephelometer data should be recorded using the Logging feature in the Data Collection module of the TSI Nephelometer software, or with any terminal emulation software. The noise check should be run for 12-24 hours to determine variability in the background values.
A program can then be run on this log file that calculates means and standard deviations for the 1-minute filtered air and zero background measurements. A Perl version of this program is included in the Appendix (6.1). As with the span gas checks, a noise check should be done periodically (at least once a year) to check that instrument background values remain low and consistent. Typical ranges of the nephelometer performance statistics for the TSI 3563 nephelometers operated by the Global Monitoring Division of NOAA/ESRL (13 instruments) are shown below.
|Parameter||Mean (Mm-1)||St. Dev. (Mm-1)|
|Filtered Air, Total Scatter (all wavelengths)||0.01-0.10||0.10-0.40|
|Filtered Air, Backward Scatter (all wavelengths)||0.01-0.05||0.07-0.30|
|Neph. Background, Total Scatter (all wavelengths)||2-8||0.02-0.12|
|Neph. Background, Backward Scatter (all wavelengths)||1-9||0.01-0.12|
Values observed that are far beyond the upper end of these ranges suggest an instrument problem; additional inspection of nephelometer is suggested.
Most users ship their TSI model 3563 nephelometers in the original wooden crate, although as the crates age it may be necessary to build a new crate or purchase an appropriate shipping container. With use the blown-in foam becomes broken, so some additional cushioning may also be required. The major criteria for fabricating a replacement shipping box for the nephelometer are:
- Protection. This is the most important criterion. The nephelometer is a rather heavy instrument with hard metal edges that can break through a flimsy shipping container. The shipping box should be made of a sturdy material; for example, wood, metal, or heavy plastic have all been used successfully. The box should have form-fitting or blown-in foam so that the instrument does not shift position in the box during transport or lifting. Cardboard and light plastic boxes should not be used because they provide a lesser degree of protection, they are easily damaged, and they require frequent replacement. Pieces of foam, newspaper, styrofoam peanuts, and other types of loose packing material should be avoided because they can allow the instrument to shift position inside the box.
- Weight and Dimensions. The wooden crates that the nephelometers are shipped from TSI in weigh approximately 61 kg (134 lbs.) when loaded with the nephelometer and accessory kit. If new shipping containers are constructed, keep in mind that several international delivery services (e.g., FedEx) have limits of 150 lbs. (68 kg) for standard air freight service. Larger packages are considerably more expensive to ship.
Finally, when shipping a TSI nephelometer make sure that the inlet and outlet are tightly sealed. This will eliminate the possibility of dust, packing debris, insects, etc., getting into the nephelometer and minimize the need for taking apart the instrument for cleaning. Also, it is wise to make sure the top and bottom covers and the PMT cover are tightly secured to protect sensitive and fragile instrument components.
After an instrument is shipped, the same initial inspection and performance checks as for new instrument arrival should be performed, except that some additional maintenance and recalibration may be required. Refer to the Instruction Manual for calibration instructions. For possible maintenance required, see Routine Maintenance and Special Maintenance sections.
NOAA typically makes several modifications to the standard TSI nephelometer before deployment at field sites. These include:
- installing plastic clips to hold the circuit boards together
- replacing fan covers with a large speaker grill, and removing the metal strip down the middle of the cutout so the lamp can be changed without removing the nephelometer cover
- installation of a small solenoid valve on the 1/4-inch port fitting next to the lamp shield, to control injection of CO2 for span checks.
- installation of a second BNC-style connector on the power/communications panel so that the solenoid valve can be controlled for automated span gas checks. A coaxial cable connects the new connector to the existing BNC connector, so that the nephelometer command 'SX 5000' opens the solenoid valve and 'SX 0' closes the valve.
- cutting the nephelometer top cover lengthwise so that it can be removed without having to remove inlet and outlet plumbing
Further details are available from the authors of this report.
Maintenance procedures for the nephelometer are described in Chapter Eight of the TSI Nephelometer Instruction Manual. Most of these procedures are recommended to be done "as needed" or "periodically". Some need to be performed when the diagnostic measurements suggest it is time for maintenance. Routine maintenance procedures are relatively simple to perform and include:
- Replacement of particulate filters (yearly, more frequently at very dusty or polluted sites)
- Replacement of the fan filter (inspect yearly)
- Replacement of lamp (as needed, generally 2-3 times per year)
- Checking for instrument leaks (yearly)
- Cleaning the main cavity of the nephelometer (as needed, if instrument background goes above ~ 10 Mm-1)
- Cleaning or changing the flocked paper (when main cavity is cleaned)
- Cleaning the light pipe lens (when main cavity is cleaned)
- Calibration or replacement of the T, P, and RH sensors (check annually)
Special maintenance procedures should be performed on an "as needed" basis. These procedures are often on sensitive components of the nephelometer, so extra care should be exercised when working on these procedures. Special maintenance procedures include:
- High voltage adjustment or replacement of PMTs
- Cleaning or replacement of aged bandpass filters
- Replacement of old/scratched chopper shutter
- Replacement of EPROM chip
- Replacement of motor control microprocessor
- Replacement of chopper and backscatter shutter motors
- Adjustment/replacement of IR reflective diodes
- Cleaning of the backscatter shutter
- Replacement or realignment of the zero filter ball valve
Over time, the photon count rates of TSI nephelometers will decrease. This is usually due either to aging of the optical glass filters that pass each wavelength of light to the photomultiplier tube detectors or to the degradation in sensitivity of the PMTs themselves. If span gas calibration checks are noisy from one check to the next, it may be that too few photons are being detected to get good counting statistics. In this case, the performance of the optical filters and PMTs should be checked.
Using the TSI software in Data Collection mode, check the raw photon counts on filtered air. Under "Total Scatter" there are nine photon count rates (in Hz) to monitor. These are "CAL", "MEAS", and "DARK" for the blue, green, and red channels. CAL measures the count rate on the calibrator, or fixed brightness section of the chopper shutter. MEAS is a measurement on the open sector of the chopper, so it is essentially a measurement of the photon count rate from the Rayleigh scatter of the filtered air in the scattering volume. DARK is a measure of the dark counts; i.e., the detected count rate with no light being passed from the scattering volume of the instrument. A general strategy is to increase a PMT's voltage as long as the change increases the CAL and MEAS count rates without significantly increasing the DARK count rate.
It is difficult to recommend specific photon count rates because they are dependent on the interaction between the lamp-color filter-detector system, and lamp brightness, bandpass efficiency, and PMT sensitivity all come into play. Still, we can come up with some minimum measure of count rate acceptability based on experience in maintaining many nephelometers. These should be viewed as guidelines rather than firm thresholds of acceptability.
Under Total Scattering, the blue channel should show a count rate of at least 60-70 kHz on the CAL and at least 500 Hz on the MEAS. The green channel should be a little higher on the CAL, with a count rate of at least 100 kHz, and a minimum MEAS count rate of 500 Hz. Remember that these are recommended minimum values, and that a properly functioning nephelometer may have counts rates 2 or 3 times higher than this. If count rates are below those suggested here, statistical noise from low count rates will affect the measurements. The DARK counts for the blue and green channels are typically quite low, often below 10 Hz.
The red Total Scattering channel always shows a much higher DARK count rate than the corresponding blue or green channels, often in the several hundred Hz range. The CAL Total Scatter count rate should be at least 150 kHz, and the MEAS count rate should be at least 800 Hz.
If the nephelometer photon count rates are below the minimum acceptable guideline values listed above, there are several things to try in order to increase them. If the lamp is old, you can try replacing that. A new lamp is often a bit brighter than an old one. If the lamp is OK, then try increasing the voltage on the PMT for the low count rate channel. If increasing the voltage does not increase difference between the MEAS and DARK counts, then turn the voltage back to where it was and try replacing the PMT with a new one. Note the photon count rate before proceeding.
To replace the PMT, it is important to disconnect the power cord from the neph. Normal room-intensity light will damage an energized PMT! Then remove the PMT housing and replace the PMT. Make sure the PMT cover is well seated before reconnecting the power cord and turning on the power. If the photon count rate is significantly higher than it was before at the same voltage, then it was probably a wise move to replace the PMT.
If replacing the PMT does not increase the count rate, you can try cleaning or replacing the optical bandpass filter for that channel. Over time, these filters can become cloudy or hazy, especially in very humid environments. To remove the optical filters, again cut power to the instrument and remove the PMT housing cover. Extract the correct color filter by removing the two small screws and nylon washers holding it in place in the optical assembly. To clean an optical filter, use lens paper if available and a very clean alcohol like spectrophotometer-grade methanol. Impure alcohols will leave a deposit that will itself have to be cleaned. Very gently pull the lens paper across the optical surface to remove the alcohol. Do not apply pressure to the lens paper as the optical coating can scratch, and then light other than the correct wavelength can pass the filter. If the optical filter looks less cloudy, then the haze deposit may have been successfully removed. The only way to be sure is to replace it in the optical assembly and check the photon count rate using the same PMT voltage. If it did not increase, the only other option is to replace the suspect optical bandpass filter with a new one. Generally, some combination of cleaning and/or replacing components of the lamp-filter-detector system will increase the count rates into the acceptable range. If it does not, the nephelometer may have to be sent back to the factory for an overhaul.
We recommend replacement of an old chopper shutter, rather than cleaning. We have found through experience that it is very difficult to clean one of these shutters without leaving a dull deposit or imparting additional scratches on the reflective surface. The TSI Nephelometer Instruction Manual recommends cleaning a dirty chopper shutter with isopropyl alcohol and cotton swabs. Feel free to try this, but don’t be surprised if you end up needing a new chopper shutter anyway.
The two IR reflective diodes are used to detect when the zero valve and the chopper shutter are in the appropriate positions. The lenses for these diodes can get dirty and may need to be cleaned periodically. These diodes have been found to fail over time, so when cleaning or adjustment does not make these perform better, it is time for a new diode. If the diode needs to be replaced, note the distance between diode and shutter and try to match that with the new installation. Generally, these diodes should be 1-3 mm from the surface of the shutter to ensure position detection.
The backscatter shutter should be cleaned so that dirt or dust on the shutter does not lead to additional scattering of light from the lamp. Care should be taken not to change the orientation of the backscatter shutter (i.e., the angle at which it rotates above its base plate). If this orientation is changed, the K4 constant will change and a new calibration will be required.
Over time, the ball valve assembly can cause problems either by developing a misalignment or by becoming more difficult to turn. These problems can cause background measurements that are off by varying degrees, or in the extreme case of a ball valve that will not turn a nephelometer unable to calculate its own backgrounds. A misaligned ball valve lets ambient air into the instrument during the zero air background measurement, which obviously compromises the background measurement. This can be observed by shining a flashlight into the nephelometer inlet when the valve is supposed to be in the zero air position. Seeing a gap where air can get directly into the nephelometer confirms the problem.
A misalignment of the ball is usually caused by one or more of the four set screws that hold the couplers in place becoming loose. This permits the shaft to rotate relative to the aluminum flange that is used a positioning device. The way to correct this problem is to loosen all of the set screws so that the ball can be turned by hand. Position the ball so that it is as far open as possible; i.e., that it allows air to enter the nephelometer as efficiently as possible. Then position the flange so that its edge is directly over the IR reflective diode sensor that determines flange (and valve) position. The metal should be 1-3 mm away from the sensor. If the distance is greater than that, adjust the position of the IR reflective diode closer to the aluminum flange. Make sure to rotate the valve over 360 degrees because the flange is often tilted slightly and could move too close or far away from the diode for position detection. After aligning the ball and getting the flange in the correct position, tighten the set screws to lock the assembly in place. Make sure when the ball valve changes position during background checks that the ball is also in the proper (sealed) position at that time.
In the extreme case, an aged ball valve can become locked in position and the shaft will either break or the motor or coupling will be damaged. Replacement of the ball valve is discussed in the section "Troubleshooting and Repair".
During normal operation, the only consumable parts are the internal filters (TSI #1602051 and #1602080) and the lamp (TSI #2201111). Filters should be replaced annually, although more frequent changes may be needed in very dusty or polluted environments. Lamps should be replaced when the current exceeds 6.5 amps, which typically occurs after 4-6 months of operation.
A supply of pure (>99.9%) CO2 is required for calibration and span checks. The calibration gas must be filtered; a suitable choice is a second blue DQ filter (TSI #1602080). The calibration gas should enter the nephelometer at room temperature; a coil of copper tubing between the CO2 tank and the instrument will achieve this.
Troubleshooting and Repair
Nephelometer repairs can be tricky and in general are best left to the factory. Repairs of this type include electronic repairs, circuit board repairs, motor repairs, etc. There are a few repairs that can usually be made by a competent end user. These include:
- Replacement of broken zero filter motor, ball valve, or coupler
- Repair or replacement of ribbon cables and connectors
- Replacement of white rectangular plastic AMP connectors and attached cables
- Replacement of various chips and microprocessor on the IC cards.
If the ball valve is not turning easily, it probably needs to be replaced. This ball valve can be ordered from TSI, but can also be ordered directly from the manufacturer. The manufacturer is Georg Fischer Piping Systems. The valve is a “Ball Valve Type 346” with a 1-inch bore. See the web page at http://www.us.piping.georgefischer.com/index.cfm?6330B9B99D5F474C87D47549DE959C77. This valve is now out of production, but the manufacturer states that it will be supported with parts until 2013. If you have a broken ball and/or stem, you can simply order another ball set. The part number you will need is 161.482.877. If you need a new ball valve (including the valve body), you will need part number 161.483.943.
To replace the broken valve, loosen the 4 large hex-head bolts that secure the valve and inlet housing to the nephelometer body. Remove the broken valve, inlet housing, and HEPA filter. Remove the coupling and flange from the shaft of the broken valve and install it on the shaft of the new valve. Make sure to align the set screws with the groove in the shaft so that the ball position will be correct. Place the new ball valve in position, making sure that the couplers fit together and that the flange is close to the IR reflective diode sensor. Tighten the four hex-head bolts down to secure the ball valve. CAUTION: The ball valve body has o-ring seals at each end, so the bolts do not have to be tightened really tight. The o-rings have to be compressed, but over-tightening the bolts can impede the turning of the ball in the valve.
Replacement of the zero filter motor assembly should be straightforward – just a one-for-one replacement. Again, make sure that the couplers fit together and that the ball is aligned after the replacement.
If chips on the IC boards are suspect, they are fairly easy to remove, reseat, or replace. Simple care is required so that pins are not damaged. If corrosion is detected on the pins, a very light abrasive like emory paper or a pencil eraser can be usd to improve the contacts. The microprocessor pins are very small and probably can not be cleaned successfully with an abrasive. Special corrosion remover liquids must be used to clean the pins on the microprocessor, and special care is needed in the re-insertion of the microprocessor into its socket.
A description of tools and equipment needed for nephelometer maintenance is given in the instrument operator's manual.
Manufacturer's Operating Manual
A detailed operating manual is supplied with each instrument. An electronic copy is available at http://www.tsi.com/particledocs/3563-Integrating-Nephelometer-1933563g.pdf. Login credentials (provided with permission from TSI) for accessing this page are:
- username: tsi-particle
- password: W3bP4rt1cle
International and National Procedures
WMO, Aerosol measurement procedures, guidelines and recommendations, WMO/GAW Report No. 153, WMO TD No. 1178, World Meteorological Organization, Geneva, 2003. Available at ftp://ftp.wmo.int/Documents/PublicWeb/arep/gaw/gaw153.pdf.
Anderson, T.A., and J.A. Ogren, Determining aerosol radiative properties using the TSI 3563 integrating nephelometer, Aerosol Sci. Tech, 29, 57-69, 1998. Available at http://www.informaworld.com/openurl?genre=article&issn=0278-6826&volume=29&issue=1&spage=57
Anderson, T.L., D.S. Covert, S.F. Marshall, A.P. Waggoner, R.J. Charlson, M.L. Laucks, J.A. Ogren, R. Caldow, R. Holm, F. Quant, G. Sem, A. Wiedensohler, N.A. Ahlquist, and T.S. Bates, Performance characteristics of a high-sensitivity, three-wavelength, total scatter/backscatter nephelometer, J. Atmos. Oceanic Technol., 13, 967-986, 1996. Available at http://ams.allenpress.com/perlserv/?request=res-loc&uri=urn%3Aap%3Apdf%3Adoi%3A10.1175%2F1520-0426%281996%29013%3C0967%3APCOAHS%3E2.0.CO%3B2
Heintzenberg, J., and R.J. Charlson, Design and applications of the integrating nephelometer: a review, J. Atmos. Oceanic Technol., 13(5), 987-1000, 1996. Available at http://ams.allenpress.com/perlserv/?request=res-loc&uri=urn%3Aap%3Apdf%3Adoi%3A10.1175%2F1520-0426%281996%29013%3C0987%3ADAAOTI%3E2.0.CO%3B2
Evaluation of nephelometer noise levels from overnight zero-air runs
Instrument performance can be assessed by running for 12-2 hours with a HEPA filter on the inlet to the nephelometer. This Perl program will calculate and display the means and standard deviations of the filtered air and zero background measurements for all six channels. The following configuration commands should be issued prior to the run:
STB61 (sufficient for a high flow rate like 30 lpm, should be longer for lower flow rates)
Span check algorithm for TSI 3563 Nephelometer
In a span gas check, the scattering coefficients of a low span gas (typically filtered air) and a high span gas (for example, CO2) are measured under instrument conditions of temperature and pressure. The results are used to derive the measured scattering coefficient of CO2 under conditions of standard temperature and pressure (STP; 273.15K and 1013.25 mb). The measured value of scattering by pure CO2 is compared with the published value [Anderson et al., 1996; Anderson and Ogren, 1998] for each measurement wavelength. The mean “error” in the CO2 measurement (i.e., the difference from the CO2 target value), calculated from each of the six nephelometer channels (three wavelengths each with a total and hemispheric backscatter measurement) should be within a few percent, with no individual channel’s error being larger than 10%. If observed errors are larger than this, it suggests an instrument problem and/or a poor calibration. A span check algorithm is provided in Appendix A so that users can perform these calculations. As discussed below, span gas checks should occur at regular intervals (e.g., weekly to monthly) so that instrument performance can be tracked over time.
Span checks that show large negative values are often caused by CO2 either not entering the nephelometer as expected or not staying inside the instrument. If the CO2 is delivered under elevated pressure, hoses can be blown off fittings inside the nephelometer cover. Check to make sure no tubes have been disconnected or ruptured and that CO2 is in fact flowing through the nephelometer. Since the CO2 measurement is made relative to the measurement of filtered air, large negative errors will also be encountered if the filtered air measurement is compromised. This can happen if the zero filter ball valve is not completely sealing off the inlet and directing all air through the heap filter. If this turns out to be the case, either adjust the ball valve so that it completely seals off the inlet, or else replace it if necessary.
STB61 (sufficient for a high flow rate like 30 lpm, should be longer for lower flow rates)
- Flush with air for 3-5 minutes at ~ 30 lpm
- Turn off blower, close off output, restrict input if possible.
- Flush with CO2 for 10 minutes at ~ 5 lpm
- Measure with CO2 for 5 minutes at ~ 5 lpm
- Record average values during CO2 measurement
- Open input and output fully, turn on blower
- Flush with air for 3-5 minutes at ~ 30 lpm
- Measure with air for 10 minutes at ~ 30 lpm
- Record average values during air measurement
- Perform a zero
Average values of the following nephelometer parameters should be recorded for the CO2 and AIR measurements. Separate values are recorded for the blue, green, and red channels [λ] in most cases.
|Photon Count Records (B, G, R):|
|NTCAL[λ]:||photon counts from calibrator (total scatter)|
|NTMEAS[λ]:||photon counts from measure (total scatter)|
|NTDARK[λ]:||photon counts from dark (total scatter)|
|REVT:||revolutions of chopper for total scatter measurement|
|NBCAL[λ]:||photon counts from calibrator (back scatter)|
|NBMEAS[λ]:||photon counts from measure (back scatter)|
|NBDARK[λ]:||photon counts from dark (back scatter)|
|REVB:||revolutions of chopper for backscatter measurement|
|Data Records (D):|
|BSP[λ]:||total scattering coefficient (m-1)|
|BBSP[λ]:||back scattering coefficient (m-1)|
|Auxiliary Status Records (Y):|
|PRES:||barometric pressure (hPa)|
|TEMP:||sample temperature (K)|
|T-IN:||inlet temperature (K)|
|RH:||relative humidity (percent)|
|VLAMP:||lamp voltage (V)|
|ALAMP:||lamp current (A)|
The calculations use the following constants:
Chopper rotation rate = 22.994 revolutions per second
Chopper gate widths = (40, 60, 140) degrees for (calibrate, dark, signal) sections
Standard temperature and pressure:
- T_STP = 273.15 K
- P_STP = 1013.25 hPa
Rayleigh scattering coefficient of air at STP:
- BSGAIR[λ] = (27.89, 12.26, 4.605) Mm-1 for (450, 550, 700) nm wavelength
- BBSGAIR[λ] = BSGAIR[λ] / 2
Rayleigh scattering coefficient of CO2, relative to air:
- RAYCO2 = 2.61
Rayleigh scattering coefficient of CO2 at STP:
- BSGCO2TRUE[λ] = BSGAIR[λ] * RAYCO2
- BBSGCO2TRUE[λ] = BSGCO2TRUE[λ] / 2
Calculate average gas density and lamp power:
- DENAIR = PRES[AIR] / TEMP[AIR] *273.15 / 1013.25
- DENCO2 = PRES[CO2] / TEMP[CO2] *273.15 / 1013.25
- POWER = VLAMP * ALAMP
Convert photon counts to count rates in Hz (eq. 7-15 in TSI manual), for CO2 and AIR measurements separately:
- HZTCAL[λ] = NTCAL[λ] * (360/40) * 22.994 / REVT
- HZTMEAS[λ] = NTMEAS[λ] * (360/140) * 22.994 / REVT
- HZTDARK[λ] = NTDARK[λ] * (360/60) * 22.994 / REVT
- HZBCAL[λ] = NBCAL[λ] * (360/40) * 22.994 / REVB
- HZBMEAS[λ] = NBMEAS[λ] * (360/140) * 22.994 / REVB
- HZBDARK[λ] = NBDARK[λ] * (360/60) * 22.994 / REVB
Don't bother with dead time correction (eq. 7-16 in TSI manual), because count rates on CO2 and air are too low for dead time to matter.
Calculate CO2 Rayleigh scattering at STP, as measured by nephelometer:
- BSGCO2[λ] = BSPCO2[λ] / DENCO2 - BSPAIR[λ] / DENAIR + BSGAIR[λ]
- BBSGCO2[λ] = BBSPCO2[λ] / DENCO2 - BBSPAIR[λ] / DENAIR + BSGAIR[λ]/2
Calculate percentage error in measured CO2 Rayleigh scattering:
- ERRTS[λ] = (BSGCO2[λ] / BSGCO2TRUE[λ] - 1) * 100
- ERRBS[λ] = (BBSGCO2[λ] / BBSGCO2TRUE[λ] - 1) * 100
Calculate nephelometer sensitivity factor, defined as the photon count rate (Hz) attributable to Rayleigh scattering by air at STP:
- SENSTS[λ] = ( (HZTMEASCO2[λ] - HZTDARKCO2[λ]) / DENCO2
- - (HZTMEASAIR[λ] - HZTDARKAIR[λ]) / DENAIR )
- / (RAYCO2 - 1)
- SENSBS[λ] = ( (HZBMEASCO2[λ] - HZBDARKCO2[λ]) / DENCO2
- - (HZBMEASAIR[λ] - HZBDARKAIR[λ]) / DENAIR )
- / (RAYCO2 - 1)
Absolute values of ERRTS[λ] and ERRBS[λ] larger than a few percent indicate a potential problem with the nephelometer or with the calibration parameters stored within the nephelometer. If larger errors are encountered, the span check should be repeated. If the errors persist, the full calibration procedure recommended by TSI should be performed.
Long-term trends in SENSTS[λ] and SENSBS[λ] should be monitored for degradation of phototube sensitivity.
Document authors and history
Patrick J. Sheridan and John A. Ogren
National Oceanic and Atmospheric Administration (NOAA)
Earth System Research Laboratory (ESRL)
Global Monitoring Division (GMD)
Boulder, CO, 80305, USA
2006-10-27 original version
2007-12-02 add Urs Baltensperger comments
2009-08-06 format for wiki [JAO]
2010-03-01 add content