The correct functioning of HEPA filters is normally tested at time of installation, as well as part of the routine maintenance program. The frequency of periodic testing varies with the cleanroom production needs. Testing is necessary, since HEPA filters can become partially or totally blocked, or the air supply can be degraded. A HEPA filter is of little value if it is not performing its task, and can cause unforeseen contamination. Existing methods of testing include the measurement of the airflow at multiple points below the filter, usually 8 points, using a single point anemometer (hot wire, mechanical vane). Typically 8 separate readings have to be taken, on a 2 X 4 filter. The results are separately recorded either on paper (and then manually entered into a PC) or via a direct download to a PC, which has to be attached to the anemometer during the testing.
The single point method, while it does obtain multiple readings and provides important information on the distribution of airflow beneath the filter, is cumbersome, very time-consuming, and prone to operator error. Furthermore, the readings are not truly "real-time", since the 8 points are measured at different times.
The other methods include a "pitot grid" or "pressure grid", where multiple readings are "averaged" pneumatically in a plenum, and then measured with a differential pressure sensor. Such a system can be directionally sensitive, resulting in significant errors. In addition, it does not give any information on the laminar distribution of airflow through the filter. In fact, it can cause serious mathematical errors. A "pitot" or "pressure" grid method of airflow measurement reads the dynamic pressure (also known as the velocity pressure), which is in reality the SQUARE of the velocity. In such a grid, an average is taken of the squares. This average is measured and then a square root is taken, to provide an airflow reading. This can result in significant errors. The following example illustrates the problem.
Using the Quattro Flow Velocity Analyzer
Supposing we measure airflow at 8 points around a HEPA filter with the following results in fpm: |
| 60 | 70 | 80 | 90 | 100 | 110 | 120 | 130 |
| The average velocity is 760 divided by 8 = 95 fpm. |
| |
Using a Pitot Tube Grid
The results are as follows in velocity (fpm) squared: |
| 3600 | 4900 | 6400 | 8100 | 10000 | 12100 | 14400 | 16900 |
Average (Velocity)2 is 76400 divided by 8 = 9550 (fpm)2
The Average Velocity (Square Root) is 97.72 fpm |
| |
| The resulting error is approximately 2.8%, which must be added to the instrument error. This instrument error combined with the velocity square error can create inaccurate readings. |
The Quattro Flow Velocity Analyzer is designed to address these issues, while providing the user with the flexibility of using it for multi-point monitoring of airflow elsewhere in the cleanroom. Four individual point readings are taken at a time, and the averages computed, all in real-time. Readings are stored in memory until deleted by the user. Finally, the Quattro Flow Velocity Analyzer saves a huge amount of time. It can do all this in a fraction of the time compared to other methods, in many cases as little as one-fifth of the time. With facilities containing many different cleanrooms, each with hundreds of filters, the economics alone are compelling.
