The Environmental Consultancy HOME PAGE

Pump, rotameter, cassette or Pump, cassette, rotameter: Does it matter?

  Andrew F. Oberta, MPH, CIH
The Environmental Consultancy
(512) 266-1368
www.asbestosguru-oberta.com

Rotameters are often used to “calibrate” air sampling pumps so the concentration of airborne fibers can be calculated after the filter is analyzed. I put “calibrate” in quotes because a rotameter gives an indication of flow rate rather than the direct measurement one gets with a primary standard such as a bubblemeter. Rotameters are easy to use in the field and are sometimes built into the sampling pump itself. They can also be misleading if used incorrectly.

An article in the NAC Journal in 1989 (1) discussed the proper placement of the rotameter relative to the pump and filter cassette when measuring airflow with high-volume pumps. The authors concluded that placing the rotameter between the cassette and pump results in an erroneous reading that is higher than the true flow rate as measured with a primary standard, and that the proper placement of the rotameter is at the inlet to the cassette. I have recently conducted my own testing and am able to confirm their conclusions with data that matches a theoretical explanation. This process included measurements of the pressure drop across the filters, with results that surprised me and which I have not been able to explain.

Rotameter basics

How does a rotameter work? The type most often used for asbestos sampling, shown in Figure 1, is a block of clear acrylic with a hole drilled down the middle. (Click on the picture for an enlarged view.) Fittings at the top and bottom connect to the ends of this hole, which is slightly tapered and larger at the top. In the hole rests a ball that is sometimes called a “float.” The term is misleading because the ball is much heavier than air and doesn't float like a boat or balloon. Buoyancy has nothing to do with it: velocity pressure moves the ball. Air in motion exerts a pressure proportional to the square of its velocity, and as the flow rate increases, so does the velocity (and pressure) of the air pushing the ball up the hole.

The markings on the face of the rotameter in liters per minute (LPM) or cubic feet per minute (CFM) may be fairly close to the actual flow rate under standard atmospheric conditions. It is common practice, however, to calibrate a rotameter against a primary standard to produce a “calibration curve” from which actual flow rate can be read from the observed position of the ball. Figure 2 is an example of a calibration curve.

Figure 1. Rotameter
(cm scale added)

Figure 2 . Typical rotameter calibration curve

Figure 3. Rotameter at inlet position

 

Rotameter placement

Figure 3 shows the rotameter “upstream” of the cassette, measuring airflow as it goes into the cassette. We will call this, as Hillmann and Gladstone did in the NAC Journal, the “inlet” configuration. This configuration is often used in the field for pre-and post-sampling readings, and requires that the cap be placed on the cassette. Some worry that this departure from the “open-face” configuration used for asbestos sampling introduces an error due to the pressure drop through the hole in the cap (or Luer fitting). This pressure drop, however, is insignificant compared to that through the filter(s) in the cassette.

Figure 4 shows what Hillmann and Gladstone called the “in-line” configuration, where the rotameter is placed between the pump and cassette, and we will stay with that terminology. This is the configuration of sampling pumps with built-in rotameters, and has the perceived but irrelevant advantage of allowing airflow to be measured with the cap off the cassette (Figure 5). Disturbingly, this configuration is implied in recently-published ASTM D7201 Standard Practice for Sampling and Counting Airborne Fibers, Including Asbestos Fibers, in the Workplace, by Phase Contrast Microscopy (with an option of Transmission Electron Microscopy). (2) This standard advises: “Calibrate each sampling pump with a representative cassette in line.” The standard does not specify the placement of the cassette relative to a rotameter or other calibration device. Errors introduced by using the “in-line” configuration as we define it here may not matter at the low flow rates used for personal sampling of worker exposure, but we will see that this is not the case at higher flow rates used for area sampling, including clearance of abatement projects.

Figure 4. Rotameter at in-line position

Figure 5. In-line calibration


Experimental results

A Gast high-volume air sampling pump with a valve for flow control, a 20 LPM Dwyer Model VFB Flowmeter and an SKC one-liter bubblemeter were used to measure airflow over a range of 5 to 15 LPM. Pressure drop across the filter was measured with a Mityvac vacuum gauge. Tests were run with 25mm cassettes, one having a 0.8 µm MCE filter and one having a 0.45 µm MCE filter with a 5.0 µm backup filter.

The test sequence was as follows:

1. Set up the sampling train in the “inlet” configuration (Figure 6)
2. Turn on the pump and set the valve to the initial rotameter reading
3. Take two measurements of the airflow with the bubblemeter and record them
4. Turn off the pump – do not move the valve handle
5. Re-arrange the sampling train in the “in-line” configuration (Figure 7)
6. Turn on the pump and record the rotameter reading
7. Take two measurements of the airflow with the bubblemeter and record them
8. Record the reading on the vacuum gauge
9. Re-arrange the sampling train in the “inlet” configuration
10. Turn on the pump and set the valve to the next rotameter reading
11. Repeat steps 3 through 10

Figure 6. Inlet position calibration set-up

Figure 7. In-line position calibration set-up

The measurements were entered on a spreadsheet that is summarized in Tables 1 and 2. The closeness of the Qs and Qr columns shows that the actual volumetric flow rate through the sampling train is not affected by the placement of the rotameter. However, the actual flow rate with the in-line configuration is less than the reading on the rotameter, with the difference increasing as the flow rate increases. See Figures 8 and 9.

Figure 9 also shows that, despite the smaller pore size in the TEM filter, the pressure drop across it (black line) is considerably less than for the PCM filter (yellow line). Note that Figure 9 plots pressure drop against the rotameter reading, not the actual flow rate.

  Table 1. Test results for 25mm diameter cassette with 0.8 µm MCE filter

Inlet configuration

In-line configuration

Rotameter setting, Rs

Bubblemeter Time, sec

Qs, L/min

Rotameter reading, Rr

Bubblemeter Time, sec

Qr, L/min

ΔP cm Hg

ΔP in Hg

15

4.04

3.99

14.94

18.75

4.05

4.00

14.91

25.0

9.84

14

4.20

4.27

14.17

17.25

4.33

4.27

13.95

23.5

9.25

13

4.60

4.66

12.96

15.75

4.63

4.70

12.86

20.0

7.87

12

5.04

5.09

11.85

14.00

5.15

5.11

11.70

18.5

7.28

11

5.65

5.52

10.74

12.50

5.56

5.51

10.84

16.0

6.30

10

6.12

6.14

9.79

11.50

6.10

6.15

9.80

14.5

5.71

9

6.72

6.75

8.91

10.25

6.66

6.54

9.09

13.0

5.12

8

7.39

7.35

8.14

9.00

7.34

7.35

8.17

11.5

4.53

7

8.33

8.44

7.16

7.75

8.36

8.39

7.16

9.5

3.74

6

10.09

10.04

5.96

6.50

9.81

9.79

6.12

8.0

3.15

5

11.83

11.77

5.08

5.50

11.72

11.91

5.08

6.5

2.56

Table 2. Test results for 25mm diameter cassette with 0.45 : m MCE & 5.0 : m MCE filters

Inlet configuration

In-line configuration

Rotameter setting, Rs

Bubblemeter
Time, sec

Qs, L/min

Rotameter reading, Rr

Bubblemeter Time, sec

Qr, L/min

ΔP cm Hg

ΔP in Hg

15

4.03

4.13

14.71

17.00

4.14

4.11

14.55

15.5

6.10

14

4.18

4.29

14.17

16.00

4.19

4.09

14.49

14.5

5.71

13

4.79

4.71

12.63

14.50

4.53

4.52

13.26

13.0

5.12

12

4.99

4.89

12.15

13.25

4.87

4.90

12.28

11.5

4.53

11

5.43

5.42

11.06

12.00

5.35

5.51

11.05

9.5

3.74

10

6.00

6.07

9.94

10.50

6.01

6.03

9.97

8.5

3.35

9

6.51

6.51

9.22

9.50

6.50

6.61

9.15

7.5

2.95

8

7.47

7.54

7.99

8.50

7.44

7.34

8.12

6.5

2.56

7

8.61

8.54

7.00

7.25

8.58

8.59

6.99

5.0

1.97

6

9.96

10.10

5.98

6.25

9.86

10.07

6.02

4.5

1.77

5

12.12

12.13

4.95

5.00

11.89

11.91

5.04

3.5

1.38

The essential conclusion to be drawn from these data is that placing the rotameter “in-line” – between the pump and filter cassette – results in an erroneously-high indication of flow rate. If the rotameter must be placed in this position, it is imperative that a calibration curve be prepared with the rotameter and cassette in this configuration. This may be necessary, for example, with a pump that has a rotameter installed in the sampling train.

Figure 8. Test results for inlet position

Figure 9. Test results for in-line position

 

Explanation of results

Calculations were performed to test the hypothesis that the rotameter reads high in the in-line configuration because of the lower density of the air at the reduced pressure between the filter and the pump. This is analogous to correcting the flow rate for the effects of lower atmospheric pressure at an altitude greater than sea level. According to the ACGIH Industrial Ventilation Manual (3), the flow rate correction varies according to the square root of the ratio of the air density at sea level to that at altitude. Let's see how well that relationship held up during these tests.Table 3 shows these calculations for the 0.8 µm filter. Some columns are taken from Table 1.

The mass flow rate of the air is the same at all points in the sampling train, and at a rotameter reading, Rr, of 18.75 it is calculated by

W = (Qr)(ρ) = (14.91 L/min)(0.0012 gm/cm³)(1000 cm³/L) = 17.89 gm/min

The predicted rotameter reading – the volumetric flow rate Q1 that the “float” should sense at a static pressure of P1 – is a function of W and ρ1. Q1 is calculated as Q1 = 0.001 W / ρ1, where

ρ1 = 0.0012 [(P1/Po)] = 0.0012(20.08/29.92) = 0.000805 gm/cm³

At this mass flow rate,

Q1 = 0.001 W / 0.0012[(P1/Po) **½ ] = (0.001)(17.89) /(0.0012)[(20.08/29.92)]** ½ = 18.20 L/min

Table 3. Flow prediction calculations for 25mm diameter cassette with 0.8 µm MCE filter

Rotameter reading, Rr

Qr, L/min

ΔP in Hg

P1 in Hg

P1/Po

ρ1 gm/cm³

W, gm/min

Q1, L/min

Q1/Rr

18.75

14.91

9.84

20.08

0.671

0.000805

17.89

18.20

0.97

17.25

13.95

9.25

20.67

0.691

0.000829

16.74

16.79

0.97

15.75

12.86

7.87

22.05

0.737

0.000884

15.43

14.98

0.95

14.00

11.70

7.28

22.64

0.757

0.000908

14.04

13.45

0.96

12.50

10.84

6.30

23.62

0.789

0.000947

13.01

12.20

0.98

11.50

9.80

5.71

24.21

0.809

0.000971

11.76

10.89

0.95

10.25

9.09

5.12

24.80

0.829

0.000995

10.91

9.98

0.97

9.00

8.17

4.53

25.39

0.849

0.001018

9.80

8.87

0.99

7.75

7.16

3.74

26.18

0.875

0.001050

8.60

7.66

0.99

6.50

6.12

3.15

26.77

0.895

0.001074

7.35

6.47

1.00

5.50

5.08

2.56

27.36

0.914

0.001097

6.09

5.31

0.97

This prediction of the volumetric flow rate reading, Q1, matches the indicated reading, Rr, within the limits of experimental error, and the predictions for lower flow rates match equally well.

This leaves the pressure drop results to explain. Intuitively one would expect a lower pressure drop through a filter with larger (0.8µm) holes than smaller (0.45µm) holes, particularly if the latter is backed up with a 5µm pore size filter. Could it have to do with the number of holes in the filters?

Sir Arthur Eddington said: “You should never believe any experiment until it has been confirmed by theory.” In other words, don’t accept the result of an experiment if you can’t explain why it came out that way. I can't explain why the pressure drop across an 0.8µm filter is greater than across an 0.45µm filter backed up by a 5.0 µm filter and if anyone can, their views would be most welcome.

Conclusions

The conclusion by Hillmann and Gladstone published in the NAC Journal in 1989 that placing the rotameter between the filter cassette and the sampling pump results in the rotameter giving an erroneously high reading of volumetric flow rate has been confirmed by recent experiments. Further testing to explain this effect resulted in the unexpected discovery that the pressure drop across an 0.8µm pore size MCE filter is greater than that across a 0.45µm pore size MCE filter backed up by a 5µm pore size filter.


(1) Joseph C. Hillmann and Steven J. Gladstone, “Built-In Inaccuracy: The Necessity of Field Rotameter Calibration at the Inlet of Air Sampling Configurations,” NAC Journal, National Asbestos Council, Atlanta, GA. Winter, 1989 - 90. Available from Environmental Information Association, Chevy Chase, MD (www.eia-usa.org).

(2) Available from ASTM International, West Conshohocken, PA (www.astm.org)

(3) Industrial Ventilation, 17th Edition. American Conference of Governmental Industrial Hygienists, Lansing, MI. 1982.


© 2007, The Environmental Consultancy. All rights reserved.

The Environmental Consultancy HOME PAGE