Last week in Fitting & Tuning the XEDE, Part 3 we used The XEDE processor to tune the mixtures and ignition timing of the now non-hybrid Honda Insight. While the results gave improved fuel economy, the real-world fuel usage still failed to be as good as achieved when the car was working as a hybrid – the aim of the exercise.
But was such an aim realistic? When no-one else has followed the same path, it’s hard to answer that question!
But it was when mulling over the lack of major improvement that it occurred to me that the electronically controlled EGR valve was no longer operating. (In the fault condition triggered by the Honda’s defective high voltage hybrid battery, the EGR is turned off.) So why could a non-functioning EGR valve be important?
Exhaust Gas Recirculation
As described at EGR Comeback
], Exhaust Gas recirculation (EGR) can have major benefits – and to more than just exhaust emissions.
Set up correctly, EGR can reduce pumping losses, so improving part-throttle fuel economy. And that’s more than just theory: in the series starting at Tweaking the EGR, Part 1 we increased the amount of EGR in the Honda Insight and gained a small but clear urban fuel economy improvement.
So what about using the XEDE to directly control the Honda’s EGR valve? The XEDE has a high current output than can be pulse-width modulated (ie have its duty cycle varied) – so it can be used to drive a wastegate solenoid or an extra injector.
Or, I thought, an electronic EGR valve.
I contacted ChipTorque and they suggested the standard output of the XEDE would be OK driving the valve if the duty cycle was kept low – say below 50 per cent. Otherwise, a higher current board would need to be installed. (It was this high current facility that I suggested in Part 1 of this series be specified with all XEDE purchases.)
I wired the solenoid coil of the EGR valve to the XEDE and set up a map to control it, using the same resolution and axes (TPS and RPM) as used on the fuel and ignition timing maps.
Frequency and Duty Cycles
So what’s all this about frequency and duty cycles?
Digital signals are usually either on or off. This diagram shows a digital signal over time. You can see that the signal is either on or off.
An electronic fuel injector is operated with a digital signal - it is either open or closed. How often the injector is turned on and off per second is called the frequency of the signal. It is measured in Hertz.
The duty cycle of a signal refers to the proportion of available time it is switched on – in the case of the injector, what proportion of time it is open and flowing fuel. Duty cycle is expressed in per cent.
If the injector duty cycle is at 50 per cent, the injector is open for half the time. If the injector is at 75 per cent duty cycle, it is open for three-quarters of the available time. At 100 per cent duty cycle, it is open continuously, while at 0 per cent duty cycle it is continuously shut.
In the case of the Honda’s EGR valve, previous measurements had shown that it was operated by the factory ECU at a frequency of 108Hz and at duty cycles from 0 to 50 per cent.
(Incidentally, at an operating frequency of 108Hz, the EGR valve is not opening and shutting 108 times a second. Instead, because of the inertia of its moving parts, the valve hovers, with the duty cycle determining how high the pintle is in hover. This is much the same as a boost control solenoid.)
The pulsing frequency of this output can be set in the XMap software over a wide range: I was able to set it to 108Hz, the same as standard.
To ensure that the valve was in fact being controlled by the XEDE map, I started the car and at idle, dialled-in a large opening duty cycle. The engine immediately staggered: the exhaust gas was being recirculated. (EGR is not usually used at idle.)
I positioned in the car a multimeter showing the valve duty cycle and then hit the road, mapping about 50 per cent duty cycle on the EGR valve at low revs and loads.
The on-road difference was simply incredible!
At 40 or 50 or 60 km/h, it felt like the car was just loafing along - like the aero drag had suddenly halved or the road was always downhill. I couldn’t believe the difference that enabling EGR made – clearly the pumping loads at small throttle angles are large, and on this tiny and efficient engine, overcoming these loads takes significant power… power that is therefore not available to drive the wheels.
I grabbed the laptop and added more EGR, going up to 90 per cent duty cycle at light loads. And the car got better and better. To say I was excited was an understatement. Maybe that’s why I forgot about the warning from ChipTorque about not exceeding 50 per cent duty cycle…
A minute or so later the fun abruptly stopped: the multimeter read zero per cent duty cycle on the EGR output and, when I sniffed the Xede, there was an electronic burnt smell wafting from it…
In my excitement I had way exceeded the 1 amp current limit on this output and had blown the transistor. In fact, it later proved that I had also damaged the output pin on the microprocessor, requiring the unit take two trips back to ChipTorque for repair. As I said in Part 1 of this series, make sure you get the high current output fitted when you first buy the unit!
XEDE returned and repaired, I plugged it in and went EGR testing. This time, with more time to analyse what happened when I increased part-throttle EGR, the results weren’t unambiguously positive.
Certainly, with 70, 80 or 90 per cent duty cycle on the EGR valve, at light loads the car wafted along, performing fine at just tiny throttle openings. But was this more fuel-efficient? It certainly felt better but when the instantaneous fuel consumption read-out was studied, the fuel consumption didn’t seem to match the feel of the car.
Here’s the dilemma. More EGR reduces pumping losses, resulting in increased power being available at the wheels. As a result, less throttle opening is needed. But at the same time, increasing amounts of EGR has the potential to cause combustion instability, in turn resulting in less energy being extracted from each gram of fuel.
After doing back-to-back testing with the EGR enabled and disabled, it was clear that the fuel economy went backwards when running lots of EGR at light and medium loads. Even when running reduced amounts of EGR, the fuel economy results were still poorer than when not running EGR. (Remember also that at light loads I am using very lean air/fuel ratios. In the factory car, EGR does not occur in lean cruise.)
But in one area of engine operation EGR was unambiguously good. That was on the over-run – zero throttle at revs above idle. In those conditions, adding EGR reduced engine braking (because the engine was just sending exhaust gas around and around, rather than trying to suck air past the closed throttle).
This reduced engine braking meant that when approaching red traffic lights, the accelerator could be lifted very early. Down long, open-road hills the car would just roll and roll – with it slow aero drag and small tyres, being able to roll for literally kilometres at the speed limit. In these throttle-closed conditions no fuel is injected, so there are clear benefits to fuel economy.
The XEDE still needs to be used to do this EGR mapping – you can’t just use a simple switch to turn on EGR at low throttle angles. Instead EGR needs to be mapped so that at zero throttle opening and idle rpm, EGR is off. As revs build (still at zero throttle) so does EGR duty cycle. If little engine braking is required, the EGR valve duty cycle can be kept high right through the rev range (remember, at zero throttle opening). Alternatively, engine braking can be re-established at high rpm by decreasing EGR.
In the Honda the best results were gained running EGR at all zero throttle position loads above idle.
Actual EGR Flows
For a given EGR valve opening, the amount of exhaust gas that actually flows depends on the pressure differential between the exhaust and the post-throttle intake.
While no differential measurements have been made, this is most likely to be in times of high manifold vacuum – that is, when the throttle is closed but engine revs are high.
Therefore, the duty cycle with which the EGR valve is driven doesn’t necessarily reflect how much exhaust gas is being recirculated. One hundred per cent duty cycle at zero throttle, high rpm will flow a lot more EGR than 100 per cent duty cycle at the same rpm but with a high throttle angle.
So far we’ve used the XEDE to alter fuelling to:
lean-out the mixtures at very light loads, running as lean as 24:1 air/fuel ratio
maintain idle mixtures of around 14.7:1
use mixtures at as rich as 12:1 at higher rpm, full throttle
We’ve also increased ignition timing, advancing in the range of 5 – 9 degrees. (Note also that by intercepting the MAP sensor signal and reducing the load the ECU ‘sees’, the timing will have been automatically advanced, especially at light loads.)
Finally, a large amount of EGR has been activated at zero throttle angles at revs above idle.
It was at this stage that the on-road tuning was comprehensively refined.
Without any EGR occurring in any conditions except closed throttle, it was found that air/fuel ratios could be leaned-out over an even greater range of loads. Go too lean and the engine would start to surge and stagger – this occurred at air/fuel ratios leaner than 24:1. But, incredibly, quite good power and smooth engine performance could be obtained at 20:1, and at very light throttle 24:1 was completely useable.
In this tuning session air/fuel ratios were leaned-out in the higher rev, lower throttle position areas – for example, the sort of conditions encountered at 100 km/h in fifth gear on a flat road with no wind.
On the open road, where in this low power car throttle angles can vary greatly depending on even small changes in grade (and where a down-change from fifth to fourth or even third is quite common), the variation in air/fuel ratios is dramatic.
Accelerating up to speed from a standstill, the AFR (air/fuel ratio) will typically be in the low Twelves. Reach 100 km/h and start backing-off and the AFR will rapidly slide to 20:1; lift the throttle a little more and the AFR will be 24:1. In the right conditions (flat road, no wind) the car will run at 100 km/h at 24:1 AFR. But reach even a slight hill and move the throttle downwards and the AFR will immediately come back to 20:1, with more throttle bringing it back to 14.5:1. Mash the throttle to the floor, or change back a gear and accelerate hard, and the AFR will be at 12:1.
Therefore, the AFR can vary from 12:1 to 24:1 (and vice versa) in literally seconds.
Unlike a conventional car in lean cruise (though not that lean in the context of these AFRs!), the throttle response in the remapped Honda is seldom doughy – this is because with the fuelling mapped via RPM and throttle, moving the throttle downwards immediately takes you into a richer party of the map. The only time you can sport something odd occurring is when you’re holding a constant throttle and letting revs rise; then the map can then pass into a richer area (at higher rpm I run richer mixtures for the same throttle position) and the engine will suddenly pick-up a little as it reaches these richer mixtures.
So in conclusion, what to make of all these tuning changes? Firstly, it proved quite possible to run extremely lean air/fuel ratios in some driving conditions – and to transition smoothly and effectively to much richer mixtures as demanded power and response increased. As a result, fuel economy was able to be improved over what would otherwise be the case – in fact, to the point of matching the open-road fuel economy achieved over a test loop when the car was working as a full hybrid.
However, in city conditions the fuel economy (even with the XEDE tweaks) could nowhere match the hybrid car.
But I have to say that, overall, I was a bit disappointed in the outcome. Without the electric assist, 10kW is lost from the power output – a lot in such a small car. But even more importantly, the reduction in bottom-end torque is quite major. Simply, this makes the car very much harder to drive well, and also makes it much less enjoyable.
Could the XEDE do what was asked of it? Yes. But was the outcome on this car brilliant? Not really.
But a well-matched turbo would overcome those deficiencies - and the XEDE tuning ability could easily cater for a turbo….
The XEDE Processor was made available for this series at no charge. The repairs to the XEDE were also made without charge.