If you scan the screens of eBay it won’t take you long to find the marvellous, magical, resistor power upgrade trick.
For my car, a Honda Insight, you can buy a box (it’s probably just a resistor and a switch) that will give the Insight a claimed 30 more horsepower – or, if you want, 20 per cent better economy! Since the Insight is amongst the world’s most fuel-efficient cars, 20 per cent better economy is the stuff of hallucinations, while 30hp (about 22kW) represents a 40 per cent power increase – completely and utterly absurd.
The resistor trick is based on this idea: you add the resistor to the engine coolant temperature sensor circuit, or the intake air temperature circuit. This tells the ECU that the temperature is different to its actual value, and as a result, the ECU adds more fuel (or less fuel, depending on the direction of the modification), or more ignition timing or less ignition timing (again, depending on which way the mod takes the perceived temperature).
To suggest that such changes will result in enormous power gains, or huge improvements in fuel economy, is just rubbish.
But let’s not throw the baby out with the bath water. If a clear idea of what is able to be achieved is understood, such resistor changes can indeed be very effective. In fact, in bang for your buck terms, using simple resistors can be quite brilliant.
But forget crazy power or economy gains...
Intake air and coolant temp sensors use variable resistance designs – that is, they vary in their resistance to current flow with changes in temperature. Normally, they are Negative Temperature Coefficient (NTC) devices, where resistance (measured in ohms) increases as the temperature decreases.
For example, one coolant temp sensor has the following relationship between temperature and resistance:
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.
If you have an NTC sensor and you want the ECU to think that the temperature is colder, you add extra resistance in series with the sensor. If you want the ECU to think that the temperature is hotter, you add the resistance in parallel with the sensor.
In most cases it is easiest to initially use a pot wired as a variable resistor, so the two circuits look like this:
In series (to increase the resistance)....
...or parallel (to decrease the resistance).
So what engine changes are likely to result from altering sensor resistance?
The coolant temperature sensor controls almost exclusively the amount of fuel enrichment during times of cold running.
The output of the intake air temperature sensor is frequently used by the ECU to determine the final ignition timing advance.
But both summaries are simplistic: the ECU will use the inputs of sensors in more than just one way. In some cases, dozens of maps may be based on these inputs.
Doing it on the Honda Insight
The hybrid Honda Insight is in many ways a simple car – certainly, much simpler than the Toyota Prius. The engine, a 1-litre 3 cylinder, uses conventional engine management – not even electronic throttle control.
The (Australian-delivered) car is designed to run on 95 RON fuel but I normally use 98 RON. The RON value is purely a measure of the fuel’s resistance to detonation; nothing else. Higher octane fuel therefore has a higher resistance to detonation, so can tolerate a higher engine compression ratio and/or more ignition timing advance. (In a forced aspirated engine, to that list can be added increased boost pressure.)
To an extent, the ECU will automatically advance timing when running on higher octane fuel – but only to an extent. It expects 95 RON fuel, so it’s never going to advance timing to the degree it would if originally calibrated for 98 fuel.
So what, I wondered, would happen if I altered the signal the ECU saw from the intake air temp sensor? Since the lower the intake air temp, the greater is the engine’s resistance to detonation, if the ECU was convinced that the intake air temp was actually lower than it really was, it could be expected to run more ignition timing advance. That could in turn well suit the higher octane fuel.
The Honda workshop manual provides no real detail on the intake air temp sensor, but it is easily removed and tested. At about 35 degrees C the resistance was 1600 ohms, at about 20 degrees C it was 2000 ohms, and when packed briefly in ice it increased to 5000 ohms. (These figures are indicative only – I just wanted to confirm it was a Negative Temperature Coefficient sensor.)
So (to reiterate), higher resistances equal lower temperatures.
I snipped the signal feed near the sensor itself (this could have been done at the ECU but it was simpler to do it under the bonnet) and wired-in a 5 kilo-ohm pot wired as a series variable resistor (ie as shown above in the first diagram).[Note: if neither wire is connected to the sensor body ground, the pot can be inserted in either wire. If one side of the sensor wiring is earthed at the sensor, then the pot must go in the signal wire. On the Insight the signal wire is red/yellow.]
I used a 10-turn pot so that changes could be made very gradually, but a normal pot could be used if care was taken with rotation.
By turning the pot, I could make the ECU think the intake air temp was colder than it really was. I turned the pot and noted that idle speed rose slightly before then falling. This is indicative of an advance in ignition timing (what was wanted) followed by an idle speed correction. (Note that this change in rpm didn’t always occur – it depended on other parameters like engine coolant temp.)
I wound in about 3000 ohms of extra series resistance and went for a drive.
The greatest care should be taken to listen for detonation. If you don’t have an acute ear for it, do an AutoSpeed site search for some of the electronic detonation detectors we have described. Of course, high octane fuel should also be used.
On the road the Honda was clearly far more driveable. In light load driving, gear changes could be made earlier, a characteristic of increased light load torque. The earlier up-changes also suggest that in urban driving, the fuel consumption might be a little improved.
Specifically, 5th gear could be used up slight rises at 60 km/h, something the car was reluctant to do previously.
However, on a highway fuel economy test, there was no discernible change.
No detonation could be heard and no check engine light appeared.
But what was actually happening under the bonnet? The easiest way of finding that out is to use some test instruments.
Lachlan Riddel of Gold Coast tuning company ChipTorque was kind enough to make available a Snap-On diagnostics tool. This could read the generic OBD data stream, albeit at a very slow rate.
The OBD reader showed two important data – firstly, the temperature the ECU thought the intake air was, and secondly, the actual ignition timing being used.
With about 3000 ohms of extra series resistance, the OBD Reader showed that the ECU was measuring intake air temp at about 1 degree C, on this day about 25-30 degrees C lower than reality.
Test driving showed that the ignition timing being used in light load, constant throttle conditions at about 70 km/h was 4-5 degrees more advanced than standard (eg about 30 degrees rather than 25 degrees). The low update speed of the OBD stream made measuring dynamic and full-load changes impossible.
This increased timing advance matched my seat-of-the-pants judgement – but I’d wanted to make sure that massive changes weren’t being made that in turn were being retarded by the knock sensor.
In a MAP-sensed car like the Honda, the intake air temp will also help determine the mixtures. This is because the ECU uses rpm, MAP and air temp to calculate the grams/second of air that’s being inhaled. Tell the ECU that the air is colder (ie more dense) and the ECU will inject more fuel to go with it. With the air density in fact unchanged, this will result in a richer air/fuel ratio.
A MoTeC professional air/fuel ratio meter was used to measure actual on-road mixtures. As expected, when in closed loop, the air fuel ratio stayed at stoichiometric (ie about 14.7:1) and lean cruise (around 25:1) also appeared unchanged.
However, full load high rpm mixtures (where the car goes out of closed loop) were richened from about 12 – 12.5:1 (the full load mixtures vary a bit over the upper rev range) to 11.5 – 12:1, ie about half a ratio. This means that when this air/fuel ratio is being used, the car will use something like 4 per cent more fuel. However, that driving condition is a rare occurrence.
So with 3000 ohms added to the intake air temp sensor and using 98 RON fuel, the car drives better, can be changed up in gear earlier, and feels to have better part-throttle torque.
At full load the air/fuel ratios are richer, something that could be easily prevented if a throttle switch was used to short out the added resistance at large throttle openings. The extra fuel at high loads doesn’t worry me, so I’ll stick with a simple resistance. (You could also easily fit a shorting switch to allow an instant change back to lower octane fuel.)
But what happens when it really is cold? After all, the modification is effectively subtracting a lot from the real temp. Probably not much happens – the handbook suggests that -20 degrees C is the lowest the ECU will recognise.
By using a pot (rather than a fixed resistor) in the initial configuration, you can twiddle it to your heart’s content. When you have decided on the value that is required, the pot can be removed and its resistance measured with a multimeter. A fixed value resistor of the same value can then be wired in series with the temp sensor and the lot covered in heatshrink or tape. And that’s just what I did, using a 3.2 kilo-ohm ohms resistor (that actually measured closer to 2.9 kilo ohm!).
This is not a groundbreaking modification. There’s not 30 more horsepower or 20 per cent better economy. But this is better than those claims – because it works. There’s clearly improved driveability and potentially a little better urban fuel economy.
And at a cost of less than five cents (the cost of the resistor), that’s one helluva bang for your buck.