Last week ["DIY Modification of Car Electronic Systems - Part 1"] we discussed how to go about selecting the signals to modify in order that your desired performance outcome is realised. (If you haven't read that article first - go read it now!) This week we take a look at some actual electronic modification approaches that can be undertaken. They all cost nearly nothing and can be incredibly effective.
Modifying Coolant and Intake Air Temp Sensor Inputs
Potentiometers ("pots") are available very cheaply from electronics stores. Used carefully they can be very powerful modification tools, able to dramatically alter air/fuel ratios in addition to working well on other car systems. There are two ways in which they can be used - as variable resistors or as true potentiometers.
Any sensor that uses a variable resistance can be modified by the addition of a pot. The most typical of this type of sensor are the intake air temperature and coolant temperature sensors. As temperature changes, so does the resistance of the sensor.
The coolant temperature sensor controls almost exclusively the amount of fuel enrichment during times of cold running. This means that changing the output of the coolant temp sensor (ie the input from this sensor to the ECU) can cause more fuel to be injected. However, note that the extra enrichment may not be made in the same proportion throughout the load and rev range - the manufacturer doesn't expect that cold enrichment will be required at full load, 5000 rpm! But for example, if you want to smooth the idle in a cam'd car, enriching the idle mixtures will often go a long way to doing this (closed loop idle will need to be disabled).
As mentioned last week, another temp sensor that can be usefully modified is the intake air temperature sensor. The output of this sensor is frequently used by the ECU to determine the final ignition timing, and where the sensor is located after the turbo or supercharger is an especially good way of pulling back timing in forced aspiration cars. (Here's how I came to realise that: I had a high-boost turbo Daihatsu Mira which was still using the standard top-mount intercooler. Suddenly one day it started to detonate - something it hadn't done previously. I looked under the bonnet to find the intake air temp sensor plug dangling - without this input, the ECU was selecting a limp-home ignition timing setting, one which wasn't pulling enough timing off to cope with the high intake air temps actually being experienced. Fixed the plug and all was again well - the output of that sensor sure made a big difference to ignition timing!)
But before you modify the output of the temperature sensors, you should know how the sensor output works. Most temperature sensors decrease in resistance as the temp goes up. The table below shows the relationship between coolant temp and resistance in one sensor:
This sort of information is available in workshop manuals, or you can test a sensor using the sensor, a thermometer, a multimeter and a saucepan of hot water on the stove. To carry out this test, simply connect the multimeter to the sensor and place it in the water. Measure the temperature of the water and note the sensor resistance. Then heat the water, measuring the changed resistance of the sensor every 10 degrees C.
In the example shown above, the sensor has a resistance of 250 ohms at a normal coolant operating temperature of 90 degrees. If the resistance of the sensor is artificially changed so that it has a new resistance of 6000 ohms, the ECU will 'think' that the coolant is at 0 degrees, even when it is still 90 degrees! That will richen the mixtures - at least at light loads and on throttle transients. To achieve this outcome you need to increase the resistance of the sensor by 5750 ohms. A 10 kilo-ohm potentiometer can be used to give an adjustable range of extra resistance around this figure. Furthermore, adjusting the pot should create no danger to the engine - at worse a fault code might be registered and/or the Check Engine light come on.
If the opposite effect is needed - you want the ECU to believe that the temp is higher than it really is - you need to wire the resistor in parallel with the sensor. This type of wiring makes it a little more difficult to work out the value of resistor that you need to use, but it still isn't very hard. For example, assume that the air temp sensor has a resistance of 2000 ohms at 20 degrees C and 400 ohms at 80 degrees C. To make the ECU think that the intake air temp is 80 degrees when it's really only 20 degrees you need to remove 1600 ohms of resistance. But how do you work out the value of the new resistor?
The resistance of resistors in parallel is worked out by:
In this example we want a total resistance of 400 ohms but the sensor resistance is currently 2000 ohms. The equation can be re-arranged:
which gives a value of R1 (the new resistor) of 500 ohms. Doing these sums is a lot easier if the '1/x' button on a scientific calculator is used during the process.
With the installation of a 500 ohm resistor in parallel with the air temp sensor, the ECU will think that the inlet air temp is 80 degrees C when the intake air temp is really 20 degrees. This is very likely to retard the timing, but you won't know the exact effect until you've tried it.
As indicated earlier, changing the output of temperature sensors is most frequently carried out with coolant temp sensors (to change the low load air/fuel ratio), and with intake air temp sensors (to change the timing). However any sensor that uses a variable resistance can have its output changed in this way. Other sensors can include throttle position, fuel temperature, and transmission temperature. If the modification is wanted in only some situations, the changed sensor output can be easily switched in and out with a throttle-triggered microswitch or a manifold pressure switch.
Do the Maths?
Picking a 10K-ohm pot and simply trying it in each of these configurations is perhaps quicker and simpler than going through the above mathematical design process. You're very unlikely to cause any problems to anything in doing so, although you should always take things carefully and make sure that mixtures and detonation are being monitored.
Modifying Load Sensor Inputs
Changing the output of the coolant temp sensor is good for relatively small changes - but what if you want the power to dramatically change mixtures? Before discussing this approach, a few words of warning are required.
- Blowing-up an engine (especially a turbo one) developing a lot of power can result from one power run with lean mixtures.
- Mixtures should be monitored by an air/fuel ratio meter. Doing this with the car on a dyno gives the best control over what is happening.
- Start off rich and then lean things out - don't go the other way.
The reason for these warnings is that - with the following approaches - changing the mixtures from full rich to full lean anywhere in the rev range is very easy.
Firstly, you will need to examine the pin-outs of the load sensor - the airflow meter or MAP sensor. A power feed to the sensor (usually 5 volts), an earth, and a signal wire should be able to be located. In vane-type airflow meters, an air temp sensor is also located inside the meter, but you can ignore that. Whether the sensor's output voltage rises or falls with increasing engine load doesn't matter when using this circuit. However, this approach cannot be used with airflow meters or MAP sensors that have a variable frequency output.
Two 10 kilo-ohm linear designs pots are used. A pot works as a variable voltage divider, allowing the central wiper contact to be skewed to one voltage or another. One of the pots is connected between the 5 volt supply wire and earth. With this pot's wiper at the 5 volt end, that's the voltage available on the wiper (ie middle) terminal. With the wiper at the earth end, there will be no voltage available. Pot 2 connects between the wiper of Pot 1 and the signal output wire of the sensor. Moving the wiper arm of this pot (which goes to the ECU) will cause the signal to be either closer in level to the sensor output, or to the voltage provided by Pot 1.
The wiring is simple, and the use of the pots even simpler. To start the car, set both pots to their central positions - if the car won't start, make further adjustments with Pot 1. Disconnect the oxygen sensor from the ECU (so that it can't go into closed loop mode) and then use Pot 1 as a coarse mixture control and Pot 2 as fine adjustment. Start off by setting light loads. When the light load adjustment is about right, lift the power level required and re-do the fine-tuning using Pot 2. Obviously it's more important to get the high loads right than the light loads, but when the mixtures are set perfectly for high load conditions, they may not be right at light loads because of non-linearity in the engine's fuel needs when compared with its unmodified state. If the light loads look a bit too rich, don't worry - when the oxygen sensor is re-connected they'll lean out as the ECU learns its new mixtures - or you can switch the modification in and out as required.
Taking this potentiometer approach gives enormous power over the mixtures and so should be used with extreme care - you may want to use 10-turn pots to give even more sensitivity to the adjustment process. The results of this modification can be spectacularly successful - especially when the under AUD$10 cost is considered...
Switching Closed Loop
Any changes that you make to the air/fuel ratio at low loads will probably not be retained. This is because the car works in closed loop (using the feedback of the oxygen sensor) when setting its mixtures at light loads and idle. Unless you change the mixtures so far that the self-correction of the system is insufficient, the ECU will learn its way around the mods.
However, if you want richer mixtures just at idle (at light cruising loads the system is fine in an unmodified state) then switch the oxygen sensor on and off, so that the car is forced into open-loop when the modification is taking place. A microswitch placed on the throttle can be used to perform this switching. Doing this will almost certainly cause a fault code to be registered (and where fitted, the Check Engine light to come on), however the car will still run fine.
Next week: Turbo boost fuel cuts and altering one type of frequency output signal
Check Engine Lights
A point to consider when making changes to input sensors is that it can be easy to trigger the ECU's back-up or 'limp home' mode. When this occurs a Check Engine (or similar) light may be illuminated and the engine's performance limited. Entering this mode can occur if the ECU sees a sensor output well out of the range that it has been programmed to consider 'normal'.
This can occur if the sensor does not vary in output as would be expected, or if its output is unexpectedly high or low. Note that the ECU may make this judgement on the basis of other input sensors - for example, if the inputs of other sensors indicates that the car is travelling at 60 km/h at a large throttle opening, but the airflow meter signal still reflects idle conditions, the ECU may determine that one of these sensors is defective.
While modifications should usually be carried out so that the Check Engine light is not continually illuminated - or important (and real!) malfunctions will not be able to be seen when they occur - this cannot always be achieved.