Message Boards

Google FTO

Download and Distribute the Flyer!

The MIVEC Engine and Fuel Octane Levels - Part II

1995 Mitsubishi FTO GPX

September 2002

Article by:


The 6A12 powerplant in the FTO GPX is rated at 147kW. More importantly (as far as drivability is concerned), this engine has a very wide usable torque band.

Rather than opting for forced induction, Mitsubishi achieved this entirely through efficient engine design. This high compression 2.0 litre V6 provides good, usable torque right to the 8000rpm redline.

Good engine design notwithstanding, that 147kW figure was achieved using Japanese domestic 100RON fuel. The best quality fuel available at Australian pumps is 98RON. It was possible that performance might be reduced as a result.

There are various "octane booster" fuel additives on sale these days. They promise to restore lost power by raising the octane rating of your fuel, thereby improving efficiency, resisting detonation, etc. etc. Impressive gains have been noted for turbocharged vehicles such as the WRX, but would they make a difference when used in the FTO?

I had previously experimented with a cheap and reliable method of Measuring Engine Torque and Power. It seemed logical to make use of this by performing some simple tests on one of the octane booster products available.

For additional background information, see Part I of this article series, dealing with detonation, knock sensors and the practical effects of different octane level fuels.

How and What to Test?

If I tried to test different additive brands, I would find it hard to prevent traces of Brand A still being present when I wanted to test Brand B. The only testing I could realistically perform was to use a single octane booster product, with staged mixes of progressively stronger proportions.

I decided to select a product that had performed well in several Australian automotive magazine "octane booster shootouts"... Nitrous Formula Race Blend.

According to the promotional material on the web, as little as 1 millilitre of product per litre of BP Ultimate 98 fuel would bring the mixture to 100RON standard. I purchased a 300ml bottle (the "black" bottle) for around $30. In the advertised proportions, this would be quite cost-effective, with my average 40 litre fill-up costing an extra $4 in additive.

The testing procedure was simple...

  • Perform a baseline dyno run without any additive.
  • Add 1ml/litre of Nitrous Formula.
  • Perform dyno run.
  • Add an additional 1ml/litre of Nitrous Formula.
  • Perform dyno run.
  • ...and so on.

Some additional notes on the mixing procedures...

At every stage, additive quantities were carefully calculated to ensure the mixture achieved was correct with regard to the total fuel load in the tank.

The first application was poured in immediately prior to adding a good quantity of fuel at the bowser. This ensured that the recommended quantity was mixed in the recommended manner.

Subsequently, additive was pre-mixed with petrol in a 5 litre jerrycan before being poured into the tank. A top-up at the bowser mixed things up a little more. For good measure, a twisty drive through suburbia was perfomed on the way to each test run to make sure this brew was both shaken and stirred!


Here is the initial "control" result. This graph shows an average of two runs before octane booster was added...


This closely matches GPX torque/power curves I have seen from other vehicle dyno tests. It is interesting to note that peak torque is achieved at around the 5,000rpm mark, whereas Mitsubishi's original specifications listed peak torque occurring at around 6,000rpm.

With the reference test out of the way, it was time to start adding that Nitrous Formula...

Octane Booster Added - 1ml/litre Concentration

At 1ml/litre, the results were not inspiring. In fact, there was barely a point on the curve where power was up over baseline.

How precise is this? My experience to date with the Home Dyno setup is that results can vary from run to run by around 2% to 3%. If two graphs differ by 1 or 2 foot-pounds of torque here and there, I wouldn't draw any weighty conclusions from it.

However, I am able to state with total confidence that on this particular run, no performance improvement was apparent.

Other Test Notes

  • This run was performed at 9pm - one hour after the previous test run.
  • Air temperature had not changed (still about 12 degrees C).
  • The same duration warm-up drive is used prior to every run.
  • Additional fuel weight is compensated for in the Home Dyno calculations.

Octane Booster Added - 2ml/litre Concentration

Again, no difference from the baseline result.

Other Test Notes

  • This run was performed at 10pm - one hour after the previous test run.
  • Air temperature had not changed (still about 12 degrees C).

An Additional Complication

As discussed previously, the FTO engine ECU adjusts its timing so as to produce as much power as possible given the fuel it has to work with. It does so by using the Knock Sensor to detect pre-ignition of the fuel-air mixture in the cylinder. Anecdotal evidence suggests that rather than triggering a brief timing retardation in response to detonation, the FTO computer may make a persistent shift to more conservative ignition timing.

This is the reason that long-time FTO owners often suggest performing an "ECU reset" when a switch is made to high octane fuel. This process involves nothing more than removing the negative terminal from the battery, and leaving it disconnected for a duration. When reconnected, the ECU memory will have been cleared, and will have supposedly reverted to its default timing mode - tuned for Japanese 100RON fuel.

I had not yet performed an ECU reset on the vehicle being tested, but due to the complete absence of any measured change in performance to date, now looked like a very good time to do so!

ECU Reset and Re-Test - 2ml/litre Test 2

The evening's testing was at an end, and the FTO's battery was disconnected to allow the ECU to reset.

The following morning, a repeat test was performed. No additional octane booster was added.

Before we all reach for the champagne (and octane booster), I believe there is a hidden factor that accounts for the apparent improved result. I'll get to that in just a moment.

Other Test Notes

  • This run was performed at 10am.
  • Air temperature was about 15 degrees C (3 degrees warmer than the previous tests).

Evening Re-Test - 2ml/litre Test 3

Our first three dyno runs had occurred in the late evening. As the "ECU Reset" test had been performed at 10am, it made sense to make one more pass in the evening, when conditions were more attuned to the initial few tests.

The ECU was not reset again for this test...

Back to baseline again! The previous run showed some improvement, but it had disappeared. How could this be?

There are only two realistic possibilities as far as I could see...

#1 - Fuel Temperature

The air temperature during this test was 15 degrees C. This was about 3 degrees warmer than during the previous evening, so the Home Dyno software corrected slightly in favour of the most recent test. However (and it is a big however), the average Perth daily temperature cycle sees the lowest recorded temperatures around 5am or 6am, and the car was parked out in the open all night. It was a cold night, and this would have lowered the temperature of the fuel in the tank a great deal. The morning test run was the first use for the car that day (with an appropriate warm-up drive, of course).

In contrast, the previous day's tests were performed in the evening, with the car previously parked in the sun all day. It had also been driven regularly. Being an EFI system, fuel constantly circulates through a pressurised fuel rail at the top of the engine. Unused petrol (ie. most of it) returns to the tank - a little bit warmer.

It is highly likely that the fuel in the tank was significantly cooler for that last test. Can the temperature of the fuel being vapourised really have an effect on performance? The temperature of the outside air being drawn in certainly makes a difference. It is possible that the ambient fuel temperature could play a part in this equation as well...

#2 - ECU Timing Retardation Still Occurring

If the engine ECU had still detected some pre-ignition, and if it utilised persistent retarded timing modes, then the reversion to baseline would not be hard to understand.

In this scenario, let's assume that some mild pre-ignition occurred at (say) 7,300rpm during that morning test run. It had been preceded by an ECU reset, and the computer would be using timing designed for 100RON fuel. If you look back at the earlier graph, there is even a slight drop in power at this point (which is why I chose 7,300rpm for this hypothetical).

If the FTO engine management works by switching to progressively more retarded timing, it would do so at this point. If it also "remembers" its selected timing mode, then that would result in a "back-to-baseline" test result from our third dyno run.

Other Test Notes

  • This run was performed at 9pm.
  • Air temperature was about 12 degrees C.

So... What took away our small performance gain? Environmental reasons, or devilishly clever engine computer?

We might be able to find out with yet another test...

Octane Booster Added - 4ml/litre Concentration

If we were still on the receiving end of pre-ignition, and if the Engine ECU responded by switching to less advanced timing, then adding even more octane booster should give the fuel-air mixture more ability to resist detonation.

An ECU Reset was performed, and enough additional octane booster was added to bring the overall concentration to 4ml/litre. Here was the result:

For me, this shot down the pre-ignition theory...

First, I was extremely careful not to drive the vehicle very hard between the ECU reset and the dyno run. This torque curve on the graph was definitely the first time that engine ECU had seen over 5,000rpm. There was simply no opportunity for detonation to occur before this test.

Second, every test performed in the evening consistently returned a power curve on par with the original baseline. The single test that returned a higher curve was the one performed mid-morning, after a cold night outside. Coincidence, or environmental? I would suggest the latter. I certainly could not justify crediting the 2ml/litre fuel octane rating with the small torque increase on that run.

Incidentally, I graphed the average of all three tests running on 2ml/litre additive. The result was a near-perfect match with the baseline graph.

Other Test Notes

  • This run was performed at 7pm.
  • Air temperature was about 15 degrees C.


The FTO GPX runs just fine on straight 98 octane fuel. No measurable improvement in performance was discernible when using a good quality octane booster. While it may give you a good feeling to increase your RON level so it is closer to Japanese 100-octane juice, no benefit was noticable in the conditions prevalent during these tests.

It should be noted that these tests were performed in the cool air of spring. It may be a different matter in summer! In a heat wave, the mercury can hit 40+ degrees here in Perth. In those conditions, using 98RON fuel may indeed see some pre-ignition occur. If the opportunity arises, I will revisit these tests on a hot summer's day.

On the evidence collected so far, though, I won't be spending my hard-earned cash on any octane boosting fuel additive. While they may work wonders in a WRX, the FTO's naturally aspirated engine is perfectly happy to run on good quality unleaded at the pump.

Future Tests

I will of course continue to use the Home Dyno system to analyse engine performance on this car. One interesting test would be to capture a dyno run immediately following a heavy usage period (eg. performance driving events on a hot day). If detonation is occurring on straight 98 octane petrol, then it would raise its ugly head under these conditions.

It would be interesting to compare the results with the "baseline" graph seen here, to determine if our reference curve was actually a "best case" run with no ignition timing retardation...

(c) 2002, All Rights Reserved