Most people don’t have personal access to a dyno – instead they’re forced to either hire that expensive tool or do without. However, it’s possible on many cars to use the airflow meter output signal as a good guide to the performance improvements you’re making – or not making, as the case may be. If the output of the airflow meter shows that the engine is consuming more air, chances are that it is also making more power.
That information can be very valuable when you’re making tuning changes that all fall within a range that’s safe for the engine. So for example, what air/fuel ratio gives best power? That depends on the specific engine – its mechanical design, the working of other engine management systems (eg how cam timing is varied), and what modifications have been made.
In a typical naturally aspirated engine, any full-load air/fuel ratio from about 13.0 to 10:1 will be ‘safe’ – but which gives best power? It’s not sufficient to state that the answer is ‘12.5:1’, or any other number. What’s needed is some experimentation across a range of viable values.
The answer that you find may surprise you....
In this case we were experimenting with the full-load air/fuel ratio of a Toyota Prius. The Prius normally stays in closed loop (ie the oxygen sensor controlling the air/fuel ratio in a feedback system) all of the time. The result is a constant air/fuel ratio when the engine is running of 14.7:1. As we’ve covered in a previous story Altering Closed Loop Mixtures) , we’ve used the Simple Voltage Switch kit working off the airflow meter output to disconnect the oxy sensors above a certain load threshold...
... with the mixtures in this forced open loop mode set by a Digital Fuel Adjuster kit.
So, what air/fuel ratio should we set for best power? As we said, in any car that’s a problematic question but in the case of the Prius, it’s damn-near impossible to answer off the cuff. Consider for instance that the Prius engine uses the Atkinson cycle, where the expansion ratio is much greater than the compression ratio. This oddity is achieved by unique valve timing, with the engine also using adjustable intake valve timing. Nominal compression ratio is sky-high at 13.5:1! Also, since the normal air/fuel ratio is 14.7:1, wouldn’t the engine be optimised in power output for this air/fuel ratio? Finally, the Prius is difficult - if not impossible - to dyno because its hybrid petrol/electric driveline shuts down momentarily when wheelspin is detected.
In short, it looked as if using the airflow meter voltage output and doing some intensive road testing would be a very good way to get a handle on what mixtures work best in this engine.
To use this technique you’ll need a way of measuring air/fuel ratios and a multimeter. At a pinch you can get away with using a cheap LED Mixture Meter working off the car’s narrow band oxygen sensor (see Real World Air/Fuel Ratio Tuning), but preferably a professional level air/fuel ratio meter like this MoTeC unit should be used.
You’ll also need a good multimeter – one that has a peak hold facility makes things easier, and the meter should also have good resolution. Whether the airflow meter outputs a voltage or a frequency doesn’t matter – the multimeter will be able to measure either. You’ll also need a way of varying full-load air/fuel ratios – whether that’s with an interceptor, a change in fuel pressure, or another technique.
As important as the test instruments is the test regime that you’ll use. When doing this sort of road testing you should take pains to make sure that the tests are absolutely repeatable. In other words, if you do one lot of tests in one direction and another lot in the other, it’s quite likely that you’ll end up with results that aren’t very helpful. In our situation we drove around a large country road block and did all the tests while heading uphill along the same stretch of road. (In the case of the Prius that also ensured we had an identical level of battery assist each time – ie the high voltage battery was at the same starting level each run.)
We started off by setting the Digital Fuel Adjuster (DFA) map settings to zero, with the Simple Voltage Switch set so that the oxy sensors were disconnected at load site #80. In other words, when airflow meter voltage exceeded about 3.2 volts, the air/fuel ratio started going rich - as it automatically does in this car when the oxy sensor signal is lost. In fact, this configuration results in a full-load air/fuel ratio (AFR) of 11.5:1. The multimeter was installed to monitor the airflow meter output voltage and the MoTeC air/fuel ratio meter was stuck to the inside of the windscreen.
In this configuration, the peak voltage output of the airflow meter was 3.7209V. Another run was undertaken and the voltage output this time peaked at 3.7191V – just 0.04 per cent difference! (Not all the runs repeated as well as this one but it does show how close multiple runs can be.)
Changes were then made to the map tune of the DFA to provide full-load air/fuel ratios of 13.5:1, and then 12.6:1. Measurement showed that as the AFRs got richer, the peak airflow meter output increased.
The DFA was then adjusted to provide still richer mixtures – firstly 10.8:1 (again the airflow meter voltage went up!) and then a very rich 10.4:1. Incredibly, the airflow meter peak output continued to increase!
In fact, the results looked like this:
As can be seen, the airflow meter output voltage was about 0.36V higher at an AFR of 10.4:1 than it was at 13.5:1. That’s 10 per cent....
When graphed with a line of best fit, the results look like this. It’s clear that the airflow meter output voltage rises in a definable relationship with increasingly rich full-load mixtures. But where would it end? We don’t know – an AFR of 10.4:1 is rich enough for us... and at that AFR, no black smoke was evident out of the exhaust.
But perhaps the increasing airflow meter output voltage didn’t in fact correspond to an increase in power from the engine? To make sure that we weren’t barking up the wrong tree, we undertook a performance test at two different air/fuel ratios. These were rolling 60-90 km/h times, taken in hilly terrain but very repeatable.
The first performance run was done with the full-load AFR of 10.4:1. This resulted in a 6.3 second time. Then the DFA was re-mapped to provide a full-load AFR of 13.7:1. Measurement of the airflow meter output voltages at these two AFRs had shown around a 10 per cent difference in output, but would the difference be realisable on the road?
It was... the 60-90 km/h split with the leaner AFR was measurably slower at 6.5 seconds.
We could have kept on trying different air/fuel ratios, measuring airflow meter output voltages and then doing stopwatch runs. (And of course the only cost is in time and a bit of fuel!)
However, for my money, I think that the writing is clearly on the wall. This engine gives best power with really rich air/fuel ratios...
Of course, at these AFRs emissions will be hugely up over standard (although only when well away from any loads that would be encountered in the Australian emissions test cycle) and the full-throttle fuel consumption will be much increased. However, these aren’t the sort of mixtures likely to be encountered in the vast majority of normal driving – they’ll be used only when all-out performance is actually wanted.
And if I can have a noticeable increase in full-throttle performance – or, alternatively, drive gently and get awesome economy and emissions – then that’s pretty good!
And the airflow meter ‘dyno’? As was shown in this case, conventional knowledge doesn’t always show what works best in the case of specific engines. If you can be sure that the range of mixtures (or whatever other aspect you’re tuning) stays within the ballpark that’s safe for the engine, it’s a technique well worth exploring...