This section describes a system of regular checks similar to that proposed in the 1979 National Survey Advisory Handbook on Quality Control. The six-monthly tests should be duplicated by two people if possible, and in addition, the six-monthly checks should be carried out if a change of operator occurs.

Each time the site is visited, the operator should;

1. Check that the funnel is intact and properly positioned.
2. Check the tubing between the funnel and sampler is free from obstructions, kinks, items which could crush it, etc.
3. Check the sampler’s internal tubing for obstructions, kinks, moisture, loose or badly sealing joints.
4. Check that the unions to the gas meter are tight.
5. Check that the expected bubbler is active, and that the timer is showing the correct number of hours to the next changeover.
6. Ensure that the rate of bubble formation appears normal.
7. Calculate the flowrate over the past week; is it within the acceptable range of 2 m³ per day, + 10%.
8. Note any unusual occurrences which might affect results, eg. bonfires or demolition work in the area.

These should be performed at the start of pollution months April, June, August, October, December and February.

1. Examine the smoke stains obtained during the past two months. If there is any blurring, fuzzy edges, or evidence of non-centring attend to the clamps.
2. Check the reflectometer readings on your copies of the data return forms. Are any unusually high or low?
3. Pick one stain from each week in the period, and re-measure them. Ideally, this should be done by a different operator. The original and repeat readings should agree to within + 2 units for individual stains and +1 unit on average. Check that the papers were read on the correct side. Look for any overall bias between first and second readings, and if any fault is suspected contact the manufacturer of the reflectometer and arrange for it to be serviced.
4. Check the current batch of 1 vol hydrogen peroxide as follows: rinse a clean conical flask with a little of the solution. Pour in approximately 50ml, and add 3 or 4 drops of indicator. The colour should be neutral grey; with neither a pinkish or bluish tinge. If there is a hint of colour, titrate with previously checked 0.002M borate or 0.002M sulphuric acid depending on whether a pinkish-grey (acid) or bluish-grey (alkaline) colour is observed. No more than 0.2 ml of acid or borate should be needed to produce neutral grey. If not, repeat the test. If consistently outside the limit, make up a fresh batch of 1 volume, nominally pH 4.5 hydrogen peroxide, and follow the procedure in Section 6 for preparation of a correct colour control.
5. Check the accuracy of the volumetric solutions (acid and borate) using the procedure in Section 6.5. If at all possible, 5 test titrations should be done by two separate operators. The titre should be between 9.60 and 9.90 ml; if consistently outside, discard the solutions and prepare fresh batches. nb. This should also be done every time a fresh batch of solution is prepared. Besides being within the prescribed limits, the spread of the individual titres for each determination should be within 0.2 ml.

In addition to the two-monthly checks, the following should be carried out at the beginning of the pollution months April and October.

1. Advise AEA Technology Environment if there have been any changes which might mean the sampler has to be relocated in future.
2. If it is necessary, move the sampler to a warmer/cooler location for the forthcoming season, or add/remove insulation.
3. Check the sampler’s internal tubing and components for leaks or poor seals.
4. Replace any tubing which has become sticky, mouldy, discoloured, or hard.
5. Check that the inlet tubes in the Drechsel bottles are long enough, and the volume of peroxide being used is adequate, especially before the summer.
6. When carrying out the smoke stain re-measurements, make sure that any possible problems noted at an earlier stage have been dealt with.
7. If possible, it is useful to exchange a sample of smoke stains with a neighbouring Local Authority for measurement using a different reflectometer. The same criteria apply as above, ie. the original and repeat readings should agree to within + 2 units for individual stains and +1 unit on average. If not, there may be a fault with the operation or calibration with one of the reflectometers, which should be investigated.
8. Check supplies of concentrated acid/borate ampoules, concentrated peroxide, indicator etc. are not past their expiry date.
9. Prepare fresh solutions of 0.002 M sulphuric acid and obtain from AEA Technology Environment (or prepare) fresh 0.002M di-sodium tetraborate solution if needed.
10. Check the accuracy of the acid against the borate by the procedure given in Section 6.4. If possible, two operators should carry out the tests. Again, it may be useful if possible to exchange samples with a neighbouring Local Authority for testing.
11. Review the area where titration takes place. Is there adequate light? Are the work surface and walls behind the burette white, or covered with white paper or bench protector? Is a white lab coat or overall available?
12. Ensure all glassware is clean and free from contamination.

When the sampler is in use, all components preceding the pump will be under negative pressure with respect to the surrounding atmosphere. Therefore, any leaks or bad seals will allow air to leak in from the room in which the sampler is kept. If such air ingress occurs, the actual volume of "dirty" outdoor air passing through the filter and bubbler will be less than the total volume registered by the meter, causing underestimation of the ambient smoke and SO₂ concentration. The standard method for determination of black smoke concentration, ISO 9835 (1993) (equivalent to BS 1747, Part 11, 1993) states that "The leakage across the filter and valves .... should not exceed 2% of the total flowrate".

Testing a sampler for leakage is not always straightforward. The original Instruction Manual (1966) gives a method for testing individual components, using either a vacuum dessicator, or a sealed flexible container made from a plastic bag. This method is perfectly sound, however it is rarely used. Many site operators do not have access to a vacuum dessicator, and the plastic bag method has been found awkward. Also, this method does not address the problem of leaks in the connections between components, and is not practical for testing a whole sampler.

It is possible to measure the leakage between the sampler’s inlet manifold and the meter, by attaching a spare meter to the inlet, and comparing the volume registered by the two meters over a given time (eg. one hour). However, for this to work, both the spare meter and the meter in the sampler must have been accurately calibrated beforehand, against the same standard.

A quicker alternative is to use a variable-area flowmeter (also known as a "rotameter" or "float" flowmeter), to compare the flowrate into the sampler inlet manifold with the flowrate into the meter. A variable area flowmeter capable of reading upto around 1700 cc per minute (1.7 litres per minute) will be needed, together with some soft, flexible tubing (it is helpful to use tubing softer than the PVC tubing used in the sampler). Tubing of 8mm nominal bore will fit most inlet manifolds and some meter connectors; 5mm may be needed to fit the plastic connectors on some "Remus" gas meters. Some quick-release connectors will be necessary if different sizes of tubing are needed to fit the the manifold, meter inlet and flowmeter. "Jubilee" clips may be useful to secure the tubing. The method is not perfect, and is affected by the pressure drop across the various components. However, it provides a quick way of identifying samplers with serious leakage problems. Carry out the test with a clean filter in the active clamp; a dirty filter will increase the pressure drop between the manifold and meter, but the clamp may not seal properly without a filter in place. The procedure is as follows.

1. Switch off the sampler and carefully remove the inlet tube from the manifold inlet.
2. Connect one end of the soft tubing to the top of the flowmeter tube, and the other to the manifold inlet.
3. Ensure the tubing is not kinked or constricted, and switch the sampler back on. Check that it is bubbling normally. Hold the flowmeter vertically, or support it in eg. a laboratory clamp stand.
4. Read the flowmeter, and record the reading.
5. Detach the soft tubing from the manifold inlet.
6. Gently disconnect the existing PVC tubing from the inlet of the sampler’s dry gas meter.
7. Connect the top of the variable area flowmeter to the meter inlet using soft tubing. The flowmeter and gas meter inlets may require different sizes of tubing to fit (eg. 8mm and 6mm nominal bore). If so, use quick-release connectors to join them.
8. Hold the flowmeter vertically, and record the reading. This is the flowrate into the meter.
9. Subtract the inlet flowrate (which should be the lower of the two values) from the meter flowrate. Divide the resulting value by the meter flowrate. Multiply by 100 to obtain the percentage leakage.
10. Disconnect the flowmeter, reconnect the meter inlet as normal, and check that the sampler resumes normal bubbling.
11. Remember to reconnect the sampler’s inlet tube to the manifold so it resumes sampling outdoor air!

In our experience it is rare to obtain a result less than 2%, even with a sampler in good condition; 5% is a more usual figure. This difference is likely to be due to pressure drop across the components, rather than air ingress. However, if the result is greater than 5%, this indicates there is probably some air ingress, and it will be necessary to investigate the cause. Sources of leakage within the sampler include the following:

• Leaks in the manifold (fairly rare). Consult the supplier if this is suspected.
• Worn clamp surfaces; re-surface as in section 5.2 or return to supplier.
• Clamps misaligned or inadequately tightened.
• Loose or damaged clamp holders in some models.
• Filter paper not covering the entire space between the clamp surfaces.
• Bubbler head not making a good seal with bubbler bottle. Use a little silicone grease.
• Internal tubing deteriorated by age and sunlight, not sealing on clamps, bottles etc. Re-tube.
• If the sampler contains quick-release connectors, these can leak due to age and wear. If they appear at all loose, replace them. Alternatively, remove them altogether, but increase the length of tubing slightly to allow for the bubblers to be removed easily.
• Leakage within the 8-port valve, due to age and wear. Consult supplier.
• Meter connectors loose (they should be just spanner-tight), or rubber bungs being used instead of proper connectors.

Leakage can also occur through any joins in the tubing connecting the funnel to the sampler. This is much more difficult to measure. For this reason, the funnel should be connected to the sampler by one single length of tubing, without joins. For more information on re-tubing, see sections 3.1 and 4.3.