The disparity between the flow bench readings and the MAF Sensor conversion table, MLHFM, was larger than I was expecting to see even accounting for the rough calculation from kg/hr to CFM.

I checked the 5v output from the ATX power supply and observed that it was providing 5.138v, a good amount more than I would have liked given the sensitivity of the MAF sensor.  I decided that I would try using a dedicated 5v power converter to provide the MAF sensor with its 5v supply.  I had previously checked this other power supply, and due to the ability to tweak the output voltage I had been able to get 5.005v from the device.

With the new power supply in place I ran the Hitachi sensor on the flow bench again.

This time the results I obtained were:

Based on these results it’s starting to look like I’ll need to tweak the power supply to get the bench readings in-line with the stock predicted values, and then I can move on to checking an 85mm housing.

Rearranged the Hitachi MAF housing to have the airflow going through in the correct direction and also attached part of the stock airbox to try and better replicate the on vehicle airflow into the sensor.

The chart above shows the predicted values for CFM, solid blue line, based on converting kg/hr to CFM; assuming 13.076 cubic feet per pound of air.  The dots are measured readings from the flow bench.

The readings are obtained by slowing increasing airflow, noting the CFM displayed on the digital manometer and corresponding voltage on the multi-meter.

As this was the first time making a full range of readings it’s possible the procedure still needs some improvement.  I noted that with no airflow through the MAF the voltage reading was 0.05 volts.  At the other end, I could obtain voltage readings in excess of 5 volts, recording 500 CFM at 5.12 volts.

With the help of several people to gather the correct components and guide me in wiring up things properly I’ve been able to add the ability to measure the voltage reported by the MAF sensor when the MAF housing is attached to the flow bench.

During the first check out of the setup, when I finally began getting readings on the multi-meter, I noticed that the voltage reported from the sensor is not very steady, except for when the airflow is zero.  I’ll be thinking about how to minimize the affect of the variation since it appears that the voltage swing is greater than what the bench is reporting as variation in the airflow.