Precision, Performance, Confidence. ™

How to Test a Dry Gas Meter/Diaphragm Gas Meter/Totalize Volumetric Flow

Diaphragm gas meters have a self contained volume that is filled and exhausted as gas flows through the meter. This typically means they are tested in volume flow rather than mass flow. To be able to do this we must know the pressure and temperature at the meter to calculate the volume flow rate at the meter for reference data.

These types of meters typically have their own internal means of compensating for actual pressure and temperature so generally do not need to correct their readings or standardize them to a specific pressure/temperature (this can always be calculated later if actual pressure and temperature is known).

To gather the actual pressure and temperature it is ideal if the meter has internal sensors for this but they typically do not. So, next best option would be to connect sensors inline as close to the meter as possible. It is commonly accepted to use local atmospheric pressure as these type of meters typically only have very small pressure drops across them (maybe a few inWa) and vent to atmosphere. Based on the typical accuracy of these devices (often ~1% full scale) would likely be considered insignificant. For a temperature alternative, the molBox temperature reading is often used. The change in temperature of the gas from this point to the meter is often very minimal and therefore also likely insignificant.

In the meter pictured above, the gas flowing through it will cause the large dial to rotate and one revolution will for example equal 10 lm (actual liters per minute). The smaller dials then keep track of how many revolutions over time the meter has seen for coarse values. In other words, the meter just adds up flow over time or "totalizes flow" rather than showing a current flow rate and does not return to a zero point. 

Since these meters just totalize flow it can be confusing as to what flow rates to set during testing. Since there is typically a statement from the manufacturer as to what flow rates it can handle while staying within its stated accuracy, we would typically run a few test points across this range to verify the spec. For example, if the spec sheet states it has accurate readings from 5 lm to 150 lm we would run at least three points of say 5, 75, and 150 lm to validate this statement.

When running a test point it is recommended to start flowing and let the large dial rotate until it is at the 12 o'clock position so that the start point is known and would thus become your zero flow starting point. It is generally hard to stop flowing or see when directly over a primary marking so will need to interpret where the needle was when stopped so thus another reason why a long totalizing time is recommended to minimize any error reading this.  

The molbox interface, our COMPASS for Flow software, or manual data capture could potentially be used to test this type of meter. Could also run it in mass flow and convert data later.

Using the molbox

Press the Unit button and select "vlm" (volumetric). It will then go to a screen to select the actual unit (example lm = actual liters per minute). It will then prompt you if want to edit the P&T (pressure and temperature inputs), so select Yes. Options to use the either molBloc or user manual input for temperature can then be selected. The molBloc can be used if you dont think there will be a significant temperature change between the molbloc and meter that are testing or can select the user input to manually enter the temperature you get from another device monitoring it. Next it will ask to manually input the pressure at the meter. It is recommended to start flowing the first point, gather the temperature/pressure values while flowing, and then re-select the vlm unit for each point to ensure getting correct data as the pressure and temperature will likely change a little at each test point.

Now press the Display button and select the feature to "Totalize". It will prompt you to enter the amount of time to totalize over. We recommend no less than 10 minutes per test point to help ensure any differences of starting/stopping the timer or similar are insignificant. Can then press the Enter button on the moBox front panel to start the timer and it will totalize the flow. Press the Enter button again to stop and it will show the total time and total flow.

*** Avoid using A+B mode if possible, or extend totalizing time as it we have seen some extra delays.

*** The molBox clock is not calibrated. It is recommended to use a calibrated stop watch for the time ideally.

Using COMPASS for Flow

COMPASS does not have a direct feature for totalizing but there are methods to do this.  The most convenient is likely to be setting up a Test in actual units (example alm), select the appropriate device for temperature and pressure, and have each test point average the data for the recommended 10 minutes plus. This will produce an average flow rather than a totalized flow but can then multiply the average flow by the averaging time to get a totalized flow manually when the test is complete. You should check min/max values that COMPASS records and standard deviation to make sure no outliers or bad data skewed the data. Could also plot the data on a strip chart while running to look for these outliers and also see if flow might be increasing or decreasing in one direction which is not desirable.

*** Consider using an MFC as a dedicated controller to keep the flow rate well controlled to limit risk of bad data. 

*** Note that when volumetric units (like alm) are being used a separate initialization window will appear towards the end when starting the test to select the actual temperature and pressure device. This is different then say a rotameter with square root density correction where we select its temperature and pressure device on the DUT initialization screen. In other words, leave the DUT temperature selection at none as this can be confusing as to which one to select the temperature device.

*** Avoid using A+B mode if possible, or at least doubling the totalizing time as switching to get data from both channels will cause delays.

*** The averaging time is not a calibrate component. It is recommended to use a calibrated stop watch for the time ideally.

*** If any issues seen in COMPASS here are some recommendations:

      A. Make sure molBox is NOT set to display units of vlm. Select a more standard unit like                    ulm/slm from the front panel of the molbox. COMPASS does its own conversions in                        background and likely won't behave well if the molBox is already in vlm units.

     B. If using the molBloc or any other device getting temperature/pressure data automatically             this could cause issues. Try using manual entry for the temperature and pressure selections           instead if having issues. This might be even more necessary when using a sonic molbloc               with BPR set to default of Auto. 


Using manual data capture

Another method is to simply lock onto a nominal flow rate using the molbox by displaying the volumetric unit with proper temperature and pressure values entered, monitor it during testing to ensure it does not drift around enough to be significant, increase the totalizing time accordingly to help ensure any drift is not significant, and then just take the observed flow rate during the test and multiply it by the time on a stopwatch. Using an MFC to control the flow rate actively could really help in this method.


Run in mass flow

Any of the above methods could also be run with a mass flow unit to perhaps make the testing simpler. Would still need to record the actual temperature temperature and pressure values during testing so could later convert to volumetric flow. Consider this also if software issues for volumetric units are preventing you from gathering data as desired. Here is an article that discusses how to convert from mass to volume flow:




Have more questions? Submit a request


Please sign in to leave a comment.