Modification: Larger AFM	Last Updated: Saturday, January 20, 2007

This writeup was done by my dad.

We had heard that a larger AFM was made by Toyota, and used on Supras of some years, but the circuit board inside had a reverse output characteristic and it was necessary to swap boards. We finally decided to go to the local U-Pull-It yard and look. The only 2 Supras in the yard had already lost the AFM, so we decided to just wander. It soon became apparent that other Toyotas and some Mazdas had the same type AFM, and used the larger size. We wound up buying one (about $23) from a mid 90’s Celica.

Opening it up at home revealed it indeed had the ‘reverse’ PCB, so we decided to go back to the yard and see if there was another AFM with the type PCB we needed for swapping (didn’t want to destroy the original from the truck). We checked about 5 different AFM’s and were about give up but got lucky and found that a 1989 Mazda 626 had not only the correct PCB, but also the larger AFM.

CHECKING AFM FOR PCB TYPE The resistance slider in the AFM is very high resistance compared to the calibration resistors, which means it can’t be measured from the outside using an ohmmeter. However, it can be checked by using a battery and a voltmeter. Connect a battery (suggest a 9V), negative to pin 3 (gnd for the AFM) and positive to pin 4 (+12V for the AFM). Pin 1 is on the left end looking at the AFM pins and is somewhat silver colored. Measure the voltage between pin 3 and pin 6 (AFM output signal) as the flapper is moved from rest toward open – in one type (ours) the voltage increases from about 1V as the flapper is opened, and the other type (‘reverse’) the voltage decreases from about 8V.

Here is a picture of the intake side of the larger (left) and original (small) AFM’s. The bolt pattern is the same. The small one is about 1.9” X 1.9” while the larger is 1.9” X 2.5”, about a 32% increase in area. The outlet ends are round, measuring about 2 ¾ and 3 1/8 in. Mounting bolt patterns on the outlet side are different.

Of course some sort of adjustment must be made since the larger AFM should see about 30% less deflection (?) of the flapper for the same air flow. Without any changes, this would tell the ECU that air flow is 30% less than it really is, and thus the ECU would supply 30% less fuel and the engine would run really lean. This is what happened when we simply substituted the larger one – O2 voltages were very low (lean).

There are a couple adjustments inside the AFM that can correct for this. One is the spring tension on the flapper mechanism, and the other is the position of the slider assembly relative to the flapper.

Spring tension is adjusted by rotating the toothed wheel - CCW to reduce tension. The red arrow shows the place on the wheel where a little spring lever sits in a wheel tooth to latch it in position. We marked the original position of this lever on the wheel with a felt tip pen so it could be returned later if necessary (it has been rotated CCW out of view in the photo).

Slider position is adjusted by loosening the screw (green arrow) fixing the slider to the shaft. Hold the shaft (flapper valve) in position and rotate the slider as necessary. The blue arrow shows where the slider contacts the resistance element (the curved black multiple-arrow thing). Also part of the slider assembly is a rod (upper yellow arrow) which opens the fuel pump switch (lower arrow).

When first installed, the larger AFM’s slider deflection at idle was about 10 - 15 degrees less than the original AFM, causing very lean O2 voltages. However, the engine started right up and ran, but poorly, at idle. We rotated the wheel CCW until the sliders idle position was nearly the same as the old AFM, but this was far enough that at rest the fuel pump switch wouldn’t open. (Is this problem? It may only mean that the fuel pump is ON when the ignition switch is ON but the engine isn’t running. Possibly a safety issue?) We then backed off the wheel a few clicks (CW), and rotated the slider assembly CCW. This put the idle position at nearly the same as the small AFM – the engine ran much better, and the ECU was able to achieve closed loop O2 operation.

There wasn’t quite enough spring tension at rest to ensure the fuel pump switch always closed, so we bent back the post (halfway between the red & upper yellow arrowheads) holding the CW slider stop spring.

WOT AFR was richer than with the old AFM. Typically saw mid to high 12’s at low RPM. There is still a slightly lean hump in the 2500 to 4000 RPM range of around 13.0 to 13.3:1, but at least this is better than the mid 14’s earlier. Above 4000, AFR dropped steadily to about 12.0:1 at redline.

What do Spring Tension & Slider Position adjustments do? Spring tension adjusts how far the flapper (and thus the slider) move for any given airflow. Rotate the toothed wheel CW and spring tension increases, so the flapper & slider deflect less. This tells the ECU that airflow is less so the ECU injects less fuel and the engine runs leaner. This adjustment is proportional across the entire operating range of the AFM – for example10% more spring tension means 10% less indicated airflow from idle to WOT. It can be considered to be a ‘gain’ adjustment.

Slider position adjusts the position of the slider relative to the position of the flapper. Given a specific flapper position (corresponding to a specific airflow through the AFM), rotating the slider CW decreases the output signal to the ECU. This tells the ECU there is less airflow again thus resulting in leaner operation. This adjustment is not proportional – it moves the slider the same amount (degrees of rotation) at all points in the AFM operating range. This means it has a greater % affect on indicated airflow at the low end than at the high end. It can be considered to be an ‘offset’ adjustment.

Adjusting Air Fuel Ratio It would seem that to adjust AFR all one has to do is tweak the above adjustments, but unfortunately it is not quite that simple. The ECU complicates things by monitoring the O2 sensor and keeping the engine in closed loop operation anywhere between idle and moderate load. Only at high load does the ECU go open loop.

Closed loop operation means the ECU monitors the O2 sensor and continuously tweaks fuel slightly richer & leaner to keep the sensor signal cycling between < .5V and > .5V (about), thus maintaining an AFR of 14.7

So here’s a possible scenario. You head off to the dyno to do some tuning. Of course on the way over there the ECU is operating closed loop and has everything dialed in for an AFR of 14.7. A couple pulls on the dyno show that at WOT the AFR could be a little richer, so you set the spring tension wheel a few notches CCW. Another dyno pull shows AFR a bit richer, you’ve gained a few HP, so head home happy. Of course on the way home the ECU goes back into closed loop and discovers from O2 readings its running rich, and within a few short minutes has leaned fuel back out and you’re pretty much back to where you started.

It is possible to set the AFM to compensate for this, at least a little (unfortunately there’s not much range). Basically, the idea is to get the AFM to read more airflow at WOT, but not at part throttle.

AFM Schematic I traced out the circuit for the AFM (no guarantees, but think its correct). The air temperature sensor and fuel pump switch are in the AFM but not part of the circuit board.

Circuit board components are shown inside the indicated dotted lines. The board consists of two groups of components. The ‘slider element’ (the curved black 8-segment arrow-like thing) is shown as the string of 8 resistors inside the dotted lines. The slider (connected to the flapper) is the arrow and travels over the 8 resistors. The other resistors are the greenish rectangular items on the board. Silver traces connect everything together. This board has the numbers “197381-0030” and “TOCOS-9206H” printed on it.

The slider element is very high resistance – 100’s of K ohms. The other resistors are low value, quite possibly laser trimmed for calibration. The resistor values indicated are what I measured with an ohmmeter.

Click here to view the dyno run results.