Instruments4Chem

High Performance, Low Cost…

Measuring Flow Rate of a Micropump Using a Load Cell


What is a Load Cell ?

A load cell is a sensor that converts an acting load or force into an electronic signal. On this page we describe the use of a resistive load cell - basically a small bar of aluminium to which four strain gauges are attached in a wheatstone bridge configuration. By applying an excitation voltage to one pair of bridge nodes and measuring the voltage across the other pair one can use the bridge to detect mass changes as these manifest as a change in strain gauge resistances under load. The signal processing needed to do this must be capable of measuring small voltages - typically just a few tens of millivolts.

An Electronic Balance with Logging Capability

To accurately measure the flow rates delivered by MP6 micropumps, an electronic balance with a data logging capability was developed. This used an inexpensive 0-100g load cell purchased on EBay that was directly connected to the two excitation outputs and to the + and – inputs of a custom Avia Semiconductor HX711 24 bit differential ADC. The HX711 is specifically intended for use with weigh scales employing wheatstone bridge sensors.

Two Propeller pins (one input and one output) are used to provide the simple two wire interface to the HX711. Acquiring data from the HX711 is very straightforward. The data line is held high while the chip is busy and goes low when new data is available. A readout sequence can then be initiated, consisting of 25-27 clock pulses (1 usec high, 1 usec low) with the ADC data available in MSB to LSB format by polling the data line on the first 24 clocks. By providing up to 3 extra clock pulses the user can choose to use either the HX711’s A or B differential inputs and also to set the PGA gain (128 or 64) when channel A is in use. The HX711 datasheet clearly explains the three possible modes of operation.

The picture below shows the Propeller-based PCB that incorporates the HX711 circuitry. An MP6-OEM module can also be seen in the upper right part of this image. Separate mini-DIN6 connectors on the right hand edge of the PCB break out the connections to an MP6 micropump and to the load cell.

On the left edge of the PCB is a third mini-DIN6 connector to which a versatile optical detection system, or VODS, can be connected. Signals from a VODS detector (either analog or digital) can either be routed to an on-board 24 bit ADC (an LTC2440) or directly to a Propeller input pin depending on a jumper setting at J5 (bottom left corner of the PCB).
To perform mass measurements, a small metal carrier was bent up from thin aluminium sheet. A 1 cm plastic cuvette was attached by double-sided tape to this carrier, which was then placed onto the end of the load cell. The other end of the load cell was rigidly fastened via a pair of metal standoffs to a metal box.

During a run, liquid emerging from the pump outlet tube gradually collects in the cuvette. A photo of the arrangement is shown at right.

The second image below shows the system in operation with the MP6 micro pump and load cell connected to the Propeller control board.
A LabVIEWTM vi (see opposite) was developed to perform logging of data so that the linearity of mass versus time plots could be explored. The instrument was first calibrated by placing an Australian 5 cent coin on the load cell (mass = 2600 mg) and determining the conversion factor expressed in ADC counts per mg. A TARE function was also implemented so that the electronic balance could be conveniently zeroed at the commencement of a run.
To compute liquid flow rates, masses and times were recorded with the MP6 pump running at various drive settings, logging the readings into Excel files that were post-processed in Origin.

Some typical results are shown at left. Here, the pump is operating in a very low flow mode with the amplitude control set to a 30% duty cycle (giving a 1V input) but with the frequency dialled down to only 1 Hz. The datasheet for the MP6-OEM module that is actuating the piezo pump indicates that this amplitude setting corresponds to a peak-to-peak voltage of 200V.

At such a low frequency the emerging liquid stream forms individual drops that slowly grow and eventually fall into the cuvette. This behaviour is clearly visible in the staircase plot of mass versus time under these drive conditions.

The data here reveals good linearity over an ~ 90 minute run time and the slope of this plot indicates a delivered flow rate of 41 uL/min.
Additional experiments were also carried out at a range of drive frequencies (3, 5, 10, 15 and 20 Hz), resulting in the plot of flow rate versus frequency shown at right. The MP6 delivered volume (inferred from mass) showed good linearity vs. time in each of these runs and the flow rate was also found to be linear over this frequency range.

In summary, an electronic balance capable of logging mass versus time and easily capable of resolving mass differences of ~ 10 mg has been used to characterize the performance of an MP6 micropump. The MP6 pump tested performed reliably and in good agreement with the manufacturers claims.

The electronic balance was also remarkably easy to set up. Without any shielding or vibration isolation/protection from air currents the system performed extremely well. Further improvements will be explored in the future.