Measuring a MAF sensor

[2seater]

I did a search on air flow through a MAF sensor but didn't find anything useful. The reason I ask is I am comparing the flow rate through two different types of MAF sensors to get an idea of the output response to modify the tables in a chip for an ECM. What I noticed is the ratio between CFM measured by the bench and the MAF output in grams per second steadily increased as flow increased. The change was approx. 8% from 100 cfm to 320 cfm. Both the large and small MAF's I was measuring exhibited the same behavior. Perhaps it is simply the way the MAF's are calibrated or my conjecture is the air density decreases as the flow rate volume increases? Does that seem correct?

[1960FL]

2seater,

Maybe to help us understand the setup more i will ask some questions

1.) what type of MAF are you using? hot wire? vane?
2.) what are you using as the output ref from the MAF volts? millivolts? ???
3.) Have you looked at the DeltaP across the MAF, and how does this compare?

Just trying to understand how you are testing.

Rick

[2seater]

They are both GM MAF's, hot wire type that generate a frequency output. I am using a stock ECM to operate the MAF's, which includes the inlet air temperature sensor, and despite the size difference, they use the same frequency range. I am using a Fluke VOM meter with a frequency function for the MAF signal. This is an orifice type bench of my own design using plates supplied by Bruce. I have verified the calibration at four different points over the range I am using.

The testing is pretty simple. I gradually increase the air flow and record the MAF frequency at about twenty different points between approx. 50cfm and 400cfm. I do note the temperature, barometer and humidity but don't apply any correction factor. I use an approx. air density of 35grams/cu.ft. I do have the tables for the original GM MAF of kHz output vs grams per second. At 50 cfm, the cfm conversion to gr/sec is very close, within 1%, but as flow increases, the indicated mass in gr/sec steadily lags further and further behind the indicated cfm.

I thought perhaps this was caused by the depression required to flow the air but the large MAF exhibits the same ratio of lag as the smaller one and requires much less depression, 7"w.c @ 380cfm vs 19"w.c.@ 360cfm for the smaller one. The one common element is the flow measuring orifice itself?

[2seater]

Please forgive this question. I received a reply from a chip tuner I have worked with before and the numbers displayed by the MAF are artificial and are only approx. correct as a reference to see the MAF is working correctly. My flow measurements between the two are a valid reference to get in the ballpark of what to expect.

[Ravoll]

I also looked into using a MAF sensor for cfm calculations.The ones that came to mind were the HMF6 and HMF7 models.
Looked promising as the HMF6 has a digital output,and the HMF7 has both digital and analog.The problem I see is that with no air flow they both put out a huge frequency,and as soon as the littlest bit of air flow is detected ,i.e. idle speed,the frequency drops to almost nothing.As the air flow increases the frequency drops,but from the diagrams I have,only in the milli herz range.If it took bigger steps or was more linear as airflow changes,then maybe it could work.

[Tony]

There is no reason why the output of a MAF has to be linear, or conform to any particular law or curve. All the auto manufacturer is interested in is that if he makes a million of them, they all behave the same.

He then uses the fuel tables to match the non liner airflow meter to the even more non linear engine requirements.

The usual need is for fitting a lager capacity airflow meter. And the best way to do that with any hot wire system, is to remove the original hot wire sensor probe complete with all its attached electronics, and fit the whole lot into another flow meter body of slightly larger diameter. Its usually all a direct mechanical fit.

I have done this successfully several times with Nissan hot wire airflow meters after discovering that the circuit boards of different sized Nissan flow meters are calibrated very differently. All the Nissan flow meters are voltage output, and highly non linear giving very high sensitivity at low flow engine idle, and gradually reducing sensitivity up near flat out full airflow. The curve shape being very different between different flow meters.

By swapping over the original "guts" its a much less painful operation getting the system retuned.