Last issue we sorted the idle speed control. With the engine idling smoothly, it was time to hit the road and start tuning.
No start-up map
No one has ever previously fitted a MoTeC ECU (or any other programmable ECU!) to a Honda Insight, and so no off-the-shelf ‘start’ or ‘basic tune’ map is available. (If you find yourself in this situation, be aware that the tuning task has just gone up immensely.)
This lack of any available settings meant that to get the engine to even start and run, let alone drive, required a lot of experimentation.
A MoTeC Professional Lambda Meter (PLM) was used to measure air/fuel ratios. The meter’s sensor was screwed into the threaded boss on the turbo dump pipe. (This is the boss that will later hold the oxy sensor for the ECU, when the ECU is working in closed loop.)
The presence of detonation (really, the sound of the whole engine) was monitored by high quality headphones working with a small amplifier and remote-mount microphone. This system, that works extremely well, is cheap and easy to build – see Hearing Detonation.
All tuning was done on the road. There are two points to keep in mind about this. (1) I live in a rural location where there are plenty of empty roads, and (2) the Honda is a quite low-power car, even when fitted with a turbo.
I did a lot of the tuning alone (pulling off the road whenever a laptop change was needed), and some tuning with my wife or 10 year old son operating the laptop in the passenger seat.
The first surprise in tuning the Honda was that we found that negative timing advance (ie timing after top dead centre) was needed at light loads and low revs. Any more ignition advance and the engine would detonate. However, drivers of other Insights report that monitoring the OBD datastream shows that their standard cars run timing like this.
One reason that this may be the case is that the Honda’s cylinder plane is offset 14mm ‘past’ the crankshaft, changing the effective relationship of the piston to the crank.
The engine initially also felt very flat as it came onto boost – timing had to be retarded so much that the engine lacked sparkle.
I’d been tuning on 95 RON fuel that came from a country petrol station with a low turnover of this fuel; in my Honda Legend I can hear that car detonate a fraction on this fuel in summer. I then swapped to fresh 95 RON from another supplier and found that timing could be advanced as the car came onto boost, taking away some of the flatness.
Note that while I intend to run the car on 98 RON, I am tuning on the lower octane to provide a further safety margin.
New injectors and pressure reg
At this point new injectors and a new fuel pressure regulator were needed.
The standard Insight injectors are 5-hole, high impedance units that flow 152 cc/hour. Using the rule of thumb that cc/min divided by 5 gives an indication of available power (in hp), this would indicate that the standard injectors, run at standard pressure, would be good for 30hp each, or about 90hp (67kW) total.
That’s less than I am aiming for, so new injectors were required.
With some excellent guidance from the Insight Central discussion group, I selected US market 2000-2001 Toyota Camry injectors, Denso part number 23250-03010. These injectors are also high impedance and are a direct drop-in replacement.
The Camry injectors are a 12-hole design that flow 248 cc/min – 63 per cent more than the standard injectors.
Using the same rule of thumb as previously, that flow indicates that these injectors will be good for 150hp (112kW). That’s more power than I expect, but it gives plenty of capability.
Also needed was a new fuel pressure regulator – one that maintains the same fuel pressure headroom above manifold pressure, even when on boost.
I selected a Bosch 0 280 160 706 regulator, as normally fitted to a Saab 9000 turbo. This regulator is easily mounted remotely from the rail – it has external hose fittings. To plumb the regulator to the fuel rail, a low-profile Sytec fuel rail adaptor was used.
After the injector scaling factor in the MoTeC software was changed, the idle speed was as sweet as when the engine was running the standard injectors – a big relief.
With the new injectors working well, it was time to start exploring a bit more boost. A 3-port Mac valve was plumbed to the turbo wastegate and controlled by the ECU. However, as soon as boost was increased from the 30 kPa (4.3 psi) wastegate spring setting, the engine missed. In fact, it could also pop and bang as the unburned fuel went out through the exhaust.
I initially thought it a tuning problem but it occurred at points in the fuel and ignition maps where the numbers showed no dramatic change over their adjoining ones – you’d have expected to see a ‘wrong number’ in the tables for the tune to have been responsible for the problems.
I then decided to fit colder NGK LFR7AIX spark plugs. The higher the performance of the engine, generally the colder the plug fitted to it. The new plugs are a ‘7’ heat range - one range colder than standard for this Australian-delivered car. I also reduced plug gap from the standard 1.1 to 0.85mm.
With the new plugs fitted, the difference in engine behaviour was profound – not only did the missing entirely stop but the engine sound, heard through the headphones, was sweeter, especially at the difficult combination of high loads and low revs. This was presumably because the engine was further from detonation.
(That might sound odd, but if you listen carefully through amplified headphones to an engine, a harsh ‘edge’ is developed when the engine is close to detonation. Detonation itself is a different sound again.)
With the missing stopped, I could return to tuning the boost.
One thing that rapidly became clear is that the engine responds very well to boost high in the rev range. Peak torque of the standard engine is at 4800 rpm, implying that best breathing also occurs here. However, with the turbo, you’d expect to be able to move the torque peak downwards – say to 4000 rpm. That would in turn suggest that higher boost levels would be run in the midrange than the top end.
And that’s just what was able to be achieved. Lifting boost in the mid-range to about 80 kPa (11.6 psi) and then tapering it back to about 60 kPa (8.6 psi) at high revs gave good results. The timing retard required at the higher boost levels is greater than you’d expect – but then again, because of the engine design, all timing figures are less than you’d expect!
However, noticeable after boost was raised in the midrange was the ‘unwastegated’ feel that comes when boost gets higher and higher with a constant throttle angle. For example, hold a constant 50 per cent throttle from 2000 – 5000 rpm and the engine would develop a very non-linear torque output as boost swelled.
To better correlate torque with throttle position, I used a 3D table that mapped wastegate valve duty cycle against both RPM and throttle position. This gives a boost level that better matches accelerator position.
Here is a slightly simplified version of that table showing wastegate valve duty cycle (throttle position on the vertical axis and engine rpm on the horizontal):
(Note that ‘0’ doesn’t equal zero boost but instead equals ‘wastegate spring pressure boost’ - ie about 30 kPa max.)
Exhaust gas recirculation
My experience with modifying the electronically controlled EGR of the standard car (see Tweaking the EGR, Part 2 had shown that EGR was important to fuel economy. When I subsequently ran an interceptor on the car (see Fitting & Tuning the XEDE, Part 4), I found that directly controlling the EGR valve could give some very interesting results.
But what about starting from scratch with an EGR tuning strategy for programmable engine management?
Using a PWM output of the MoTeC, I found that adding EGR at small throttle angles dramatically improved the part-throttle torque – and so driveability. This is because EGR reduces pumping losses – less power is being used to drag air pass the (mostly shut) throttle because that gas is being supplied through EGR.
In the case of the Honda, the difference in running EGR over no EGR was sufficient to allow far earlier part-throttle upshifts, and the use of second gear down to very low speeds. As an example, after EGR was activated, I was able to drive up my slightly inclined driveway at 12 -14 km/h in second gear – previously, that would not have been possible.
After adding EGR, I also needed to change the action of the ‘shift up’ warning (to be covered in a later story) that indicates that an upshift should occur for best fuel economy. With EGR, that upshift could occur earlier because the engine had enough part-throttle torque to then pull the higher gear.
However, too much EGR caused bucking, and EGR also needed to be tuned very carefully to stop idling problems after a cold start (where the engine revs are set higher than for normal idle).
No EGR is used at throttle angles of more than 50 per cent.
In terms of improving around-town driveability, EGR was right up there in importance with ignition timing and fuelling – something that is not normally stated in tuning circles!
Here is the EGR PWM percentage table (throttle on the vertical axis and engine rpm on the horizontal):
With fuel, timing, boost and EGR tuned, the car was starting to drive very well.
The next steps: tuning acceleration enrichment, and making closed loop fuelling work.