When making changes to cars with engine management there are lots of ways modifying the electronic system. You can get a custom-written chip - but then you'll need to get another one done if you make more engine mods. You can buy full programmable management - but that's darn expensive and probably not needed if the modifications are fairly limited. A third approach is to use an interceptor box, a module that takes the signal coming from a sensor, changes it, and then passes it along to the ECU. The ECU then responds appropriately, giving a different injector output or ignition timing.
The Fueltronics AMFC reviewed here is designed to change the output signal of an airflow meter or MAP sensor. It costs A$450 - but see the box at the end of this feature!
What You Get
The AMFC is housed in a small, heavy black box, about 12 x 6.5 x 3cm. The box is not waterproof and so should not be mounted under the bonnet - it's designed to mount near to the ECU inside the car. Two sockets are provided on the AMFC, one for a four wire harness (that is provided) and another for a DB-9 serial cable (that isn't). An instruction sheet and software is also included.
The software consists of one screen. A table shows 20 voltage outputs from the sensor, with the current voltage output of the sensor highlighted. Percentage changes (up to ?
100 per cent) can be made to each of these 20 voltages.
How You Connect It
The AMFC requires just four connections. These are:
- Ignition-switched 12 volts
- Input from the sensor (eg the airflow meter)
- Output to the ECU
In practice, making the last two connections simply requires the cutting of the sensor signal wire, with the AMFC connected at this point.
The software is loaded into the PC and the connection to the AMFC made by the serial cable. And that's about it for setting up! Incidentally, note that if the tuning is to be carried out on a dyno you could almost get away without requiring a laptop PC, but for ease of use and on-road tuning a laptop is needed.
MAP Sensor Car
In all our testing we called upon the services of Leon Vincenzi of Adelaide's Awesome Automotive. He installed the AMFC on Mark Marchesan's Charade 1.5 litre F2 (the same car used in our Pipe Dreams intake story). During the previous story we had noticed that at full throttle the Daihatsu was running very rich (~10:1 air/fuel ratio) and so the MAP-sensed car seemed a good candidate for some leaning-out.
With the AMFC connected, the car running on the dyno, and a Dyno Dynamics/Autronic digital air/fuel ratio meter constantly reading out the mix, we were quickly able to start making changes to the MAP sensor output. Firstly, we noted that if the percentage changes being made to adjoining voltages in the AMFC chart were too great a jump, the air/fuel ratio became a bit erratic. So for example, if one load point was increased by 10 per cent in output and the next by 25 per cent, as the AMFC jumped between these two points the interpolation wasn't very smooth. This means that a 5 per cent change between adjoining load points is the maximum that can be used. In other words, the modification curve should be smooth - not changing more than 5 percentage points from voltage to voltage.
Another important point we discovered is that changes made to the sensor output at other than full throttle are influenced by the standard self-learning oxygen sensor feedback loop. This could be seen when Leon dialled-in some "lean cruise", where at high manifold vacuums the air/fuel ratio was set to be leaner than 14.7:1. However, the Daihatsu's ECU then decided that the engine was running too lean, and gradually enriched the mixture to compensate. This means that to actually enact a lean cruise mode using the AMFC, the sensor output would need to be changed to the extent that the self-learning behaviour of the ECU reached its limits, meaning that the mixtures would in fact stay lean in cruise conditions!
However, it's unlikely that people will buy the AMFC just to enact a lean cruise mode. The most common use of the device will be to change full throttle air/fuel ratios, where the ECU's closed-loop self-learning system does not work anyway. However, it should be realised that part-throttle mixture changes will be strongly influenced by the mixture correction behaviour of the ECU.
So, what about leaning-out the full-throttle air/fuel ratio? As indicated, the little Charade was running around 10:1 at full throttle - way too rich for both best power and best economy. The voltage outputs of the MAP sensor were suitably changed to give a leaner air/fuel, but then another problem reared its head. This was nothing to do with the functioning of the AMFC, but to do with how a MAP-sensed engine management system actually calculates the fuel required. A MAP sensed (ie speed density) car uses manifold pressure, engine rpm and intake air temp to calculate the engine load. This means that changing just the output of the MAP sensor isn't enough to alter the air/fuel ratio at discrete points through the load range.
Confused? - don't be! Here's the easy-to-understand result. As soon as Leon floored the throttle in the Charade, the manifold vacuum dropped to zero. That was the case whether the engine revs were 1500 or 5000. This means that changing the output voltage of the MAP sensor when zero vacuum was present altered the air/fuel ratio all the way through the rpm range by the same amount. So if the engine was running rich at the top end of the full throttle rev range (but had the correct air/fuel ratio at the bottom end), nothing could be done to correct the
high load/high rpm mixtures without also changing the high load/low rpm mixtures by the same amount.
And this is what was happening with the Charade. By changing the AMFC setting for zero vacuum Leon dropped the high rpm mixtures back from 10:1 to 12.5:1. He couldn't go any leaner (in the 13's would have been better) because the low rpm, zero vacuum mixtures were already 14.6:1! This means that going any leaner in the top end would have meant excessively lean bottom end mixtures. The blue line on the dyno graph shows that even this bottom-end full throttle mixture was starting to drop power at lower rpm, although the top end was still improving. Of course, if the air/fuel ratio had remained fairly constant through the full-throttle rev range, then the AMFC could have been successfully used to change the mixtures.
So if the AMFC has some limitations in changing the full-throttle mixtures in a MAP sensed car, what about one with an airflow meter?
Vane Airflow Meter Car
This time the car that was selected was a Toyota Sprinter running the 4A-GZE supercharged 1.6 litre engine. The engine was installed in the car by Awesome Automotive, who also tweaked the mixtures by adjusting the spring tension in the vane airflow meter. You can see a whole story on this car soon in AutoSpeed. However, Leon Vincenzi thought that while the mixtures had been adjusted, the car at full throttle could still do with a little fine-tuning.
The Sprinter's vane airflow meter has a 0-5 volts output. Initially it was thought that the car may have the (rare) 0-12 volt vane airflow meter output and so Fueltronics prepared a special AMFC that could cope with this voltage range. However, measurement of the Sprinter showed that in fact it uses the 0-5 volt range (with voltage from the airflow meter falling as load increases) and so the standard AMFC was used. The AMFC was quickly and easily wired into place and the car driven onto the chassis dyno.
As with the Charade, the AMFC showed that it could quickly change mixtures. However, a similar problem to the Charade was soon found.
In the 4A-GZE application, the output of the vane airflow meter is not proportional to all load variations. What happens is that the airflow meter snaps fully open when the throttle is floored at any rpm. The factory ECU obviously uses the airflow meter signal only in cruise and idle conditions, calculating at full load the injector pulse widths from rpm and perhaps also throttle position. With the vane airflow meter output voltage remaining constant at any full throttle engine power output from 20 to 120hp, altering the vane airflow meter output voltage had little affect on full-throttle mixtures.
Leon also tried leaning out the cruise mixtures of the Toyota (where the airflow meter signal was having an input!), but as with the Charade, the oxygen-sensing feedback loop returned the air/fuel ratio back to stoichiometric - this car very quickly compensating for the changes being made.
The problems that we found with using the AMFC are generic to any single-sensor interceptor module. Before you can use an interceptor of this type, you must be absolutely sure of what you want to achieve and how the sensor that you are modifying the output of actually works. More specifically, you need to know how the output of the sensor affects the air/fuel ratios at different engine loads and speeds.
If the signal as a whole needs to be moved up or down the load scale (for example, if fitting a larger airflow meter or larger injectors) we would expect the AMFC to work quite effectively. However, if you wish to change the full load air/fuel ratio by differing amounts at different rpm, you will be unable to do so in many cases. The AMFC will only be able to do this with a load sensor that has an output directly proportional to load across the full range of engine power. As we found out, MAP sensors (in naturally aspirated cars, anyway!) and at least some vane airflow meters do not have this characteristic. We would expect
hot wire airflow meters to work well with the AMFC, but a lack of time prevented us spending a third day on the dyno!
Changing the mixtures when the car is in closed loop mode (ie with the oxygen sensor influencing the air/fuel ratio) is also problematic. To make effective changes in this area of cruise and idle, you will need to alter the output of the sensor to an extent that exceeds the correction capability of the ECU.
While we didn't test the ability of the AMFC to overcome a turbo over-boost fuel cut generated by an airflow meter or MAP sensor, we can't see any reason why the AMFC wouldn't be quite effective at doing this.
We see the best use of the AMFC being in cars where airflow meter or injector swaps have occurred. The device is easy to install and the software simple to set up and use. Furthermore, Fueltronics has clearly made the point that they are interested in providing technical support for their product - something missing from many products of this type on the market!