Fullscreen HTML5Med Res (??MB)HTML4

The World’s Most Highly Hybrid Petrol/Elec Only a complete fruit-loop would modify a car with no less than six Electronic Control Units, a 288V battery, two electric motor/generators and a control system that frequently switches off the engine, incorporates regenerative braking and uses electronic throttle control. Well, welcome to that car – and the crazy modifier, regular SILICON CHIP contributor Julian Edgar. F or fifteen years, modifying cars has been my passion. I’ve run 21 psi boost on a 3-cylinder Daihatsu Mira, rebuilt a BMW 3.0si engine and played with cars as diverse as a V8 Lexus LS400 and a turbo Nissan Maxima V6. For the last decade or so it’s been not only a hobby but also a livelihood – over that period I’ve worked full-time for both automotive print and performance on-line magazines. But a few years ago I started getting bored. After all, there’s only so many cars fitted with huge turbos or even bigger exhausts you can feel excited about. I’d seen it all before – and it no longer thrilled. Then I came across a 16  Silicon Chip hybrid Toyota Prius at a price I couldn’t go past. A grey market Japanese import, it was the very first domestic Japanese model. Compared with the two Prius models sold new in Australia (see the Prius Models panel), it had less engine and electric power but it was still a full hybrid petrol/electric car. Here was a car I could really get my modification teeth into – literally anything I did would be cutting edge, never done by anyone else in the world. Forget ringing up the local workshop to ask what turbo size would be suitable; don’t bother joining an on-line discussion group to find out what flow injectors are fitted; siliconchip.com.au y Modified ctric Car Prius Models So what are the di fferent Prius mod els? Julian Edgar’s model, the NHW 10 (pictured at le ft), was released Japan in late 1997 in . It uses a 43kW pe trol engine and ha 30kW of electric s power available. The first model Pr ius sold outside NHW11 model wh of Japan was th e ich looks the sam e as the NHW10 has some signific but ant changes, inclu ding engine powe increased to 53kW r and electric powe r increased to 33kW The current mod el, the NHW20, . has both a new and although the body fundamentals re main the same, more significant even driveline change s. Engine power ha s increased to 57 kW and electric po er is up to 50kW. wThe NHW20 has been selling well in Australia and ar both ound the world, es pecially in the US . in a reduced intake charge. So despite the geometric compression ratio of the Prius being a sky-high 13.5:1, the cylinder pressures on the compression stroke don’t really reflect this. This approach benefits efficiency because at lower loads, the throttle is open wider for a given power output, so reducing pumping losses. To allow the degree of ‘Atkinsoning’ to be altered on the fly, the engine Electronic Control Unit (ECU) alters intake valve cam timing. The other vital ingredient in making this process work is the use of an ECU-controlled electronic throttle. In this car, the driver’s torque request often has little to do with the actual throttle angle selected by the ECU! Atkinson Cycle (sometimes called Miller Cycle) engines have low-RPM torque. But in the Prius, there’s a dirty great big 30kW AC electric motor ready to provide maximum torque at zero road speed. Two motor generators don’t worry about looking around for off-the-shelf electronic aftermarket parts. I wanted a challenge – and this was certainly it. Prius driveline The Prius uses a 1.5-litre 4-cylinder engine closely related to the engine used in the Toyota Echo/Yaris. However, it has much less power than the Echo – just 43kW. The low power output is because the engine revs to only 4000 RPM and uses what is called an Atkinson Cycle. Compared with the conventional Otto Cycle, an Atkinson Cycle engine delays the intake valve opening time, resulting siliconchip.com.au In fact, the gearbox (called the Power Split Device or PSD) actually contains two electric motor/generators. Along with the engine, these are connected to an epicyclic gear train. The engine’s output is split between the wheels and one of the generators. The generator charges the high voltage battery or alternatively, feeds the other electric motor that in turn helps drive the wheels. This electric motor can also receive power from the high-voltage battery to either assist the petrol engine or propel the car on its own. The PSD’s gear ratio is a result of the balance between the speeds of the engine, the electric motor/generators and the wheels and that depends on how much force is applied by each. This gives the effect of a continuously variable transmission. One of the electric motors also acts as a quiet and powerful starter for the engine, allowing it to be stopped and started smoothly as needed. The other generator is used to regeneratively recover energy from the car during braking and store it in the battery for later use. When the driver lifts the accelerator pedal, the engine’s fuel supply is cut off. For more on Prius technology, see the December 2001 issue of SILICON CHIP. Modifying the regenerative braking One of the first electronic modifications I performed was to increase the amount of braking regeneration. March 2006  17 To keep the Prius feeling as conventional as possible, both the regen and conventional brakes are controlled by the one brake pedal. In the first part of its travel, the brake pedal operates the regen brakes alone and as further pressure is placed on the pedal, the friction brakes also come into play. However, the regen braking in the Prius didn’t seem particularly strong. That may be because I had installed rear disc brakes and was running high-performance Kevlar brake pads with slotted discs front and back – perhaps the hydraulic brakes were doing more work than had been intended by the designers. The central colour LCD in the dash shows by means of small ‘sun’ symbols how much regen has occurred each five minutes, with each sun indicating a regen of 50 watthours. But in my driving, seeing a lot of suns was rare – so it seemed to me that if the regen could be tweaked to do a greater proportion of the braking work (especially in light braking), fuel economy would benefit and the braking would be smoother. The ABS ECU handles regen braking, sending a signal to the hybrid ECU to tell it how much regen to impose. But how does the ABS ECU know what to do? Rather than measuring brake pedal travel (which could vary with pad wear, etc), the system uses sensors to detect master cylinder pressure – the higher the master cylinder pressure, the harder the driver is pushing on the brake pedal. If the pedal pressure is low, the ECU knows that the driver wants only gentle deceleration and so instructs the hybrid ECU to apply only a small amount of regen. However, as master cylinder pressure increases, so does the amount of regen that can be automatically applied. If the amount of regen that occurs is largely dictated by the output of the master cylinder pressure sensor, why not intercept and alter this signal? That way, the ABS ECU will think that there is more master cylinder pressure than is actually occurring, so resulting in more regen being applied. Since the actual hydraulic pressure going to the brakes is unchanged, there’ll be a greater proportion of regen braking in the mix. The voltage output of the pressure sensor ranges from about 0.4-3V, rising with increasing pressure. So if a small voltage was added to this signal, the ECU should respond with more regen braking. But would the ECU detect a fault condition? The workshop Fig.1: this circuit that initially used to alter the amount of regenerative braking. A 100kW pot was wired between the output of the sensor and the regulated 5V supply, allowing the signal to be increased in voltage. To give the required fine adjustment, a multi-turn pot was used. Note: the ABS ECU has many more wiring connections than are shown in this diagram! 18  Silicon Chip A colour LCD in the middle of the dash can be configured to show economy (measured in kilometres per litre in five minute increments) and the amount of braking regeneration that has occurred. Here it can be seen that over the last 30 minutes economy has never dropped below 20 kilometres/litre (that’s 5 litres/100 km) and that 150 watt/hours of regeneration has occurred. Being driven enthusiastically and with some hard country road hillclimbing thrown in, economy of the modified car averages in the high fives in litres/100 km. manual states that a fault will be detected if the voltage from the sensor is outside the range of 0.14-4.4V, or if it is outside a certain ratio to its nominally 5V supply voltage. Further, the latter is checked when the brake switch is off (ie, brake pedal is lifted). In other words, the voltage needs to be within a certain range and in some cases this is checked with the brake pedal not activated. Fig.1 shows the circuit that was initially used. A 100kW multi-turn pot was wired between the output of the sensor and the regulated 5V supply rail, with the wiper of the pot connected to the original ECU input. However, this circuit did not take into account the selfchecking by the ECU of the input signal with the brakes not applied. Fig.2 shows how a relay was wired into the circuit so that the pot was bypassed whenever the brake pedal was Fig.2: the circuit shown in Fig.1 did not take into account the self-checking by the ECU of the input signal without the brakes applied. This diagram shows how a relay was wired into the circuit so that the pot was bypassed whenever the brake pedal was released. siliconchip.com.au released. At very light pedal brake pressure, this modification has caused the voltage at the ECU input to be lifted from 1V to about 1.15V. This results in clearly stronger deceleration when the pedal is lightly pressed and much stronger regen than normal as the pedal is pressed harder. At high braking efforts, the behaviour of the car is near standard – it’s in very light braking where there’s a clear difference. And the amount of regen now occurring? The regen braking is clearly doing much more of the braking work. This can be both seen in the display of watt-hours regenerated (there are more suns appearing) and also in the feel of the car. The regen braking is smooth and effortless, slowing the car substantially before the brake pedal is moved a little further to activate the hydraulic brakes and bring the car to a halt. It’s hard to assess overall urban fuel consumption, but in some 5-minute increments, it has improved by 30%. Intercooled turbo On the road, the Prius doesn’t feel particularly slow, despite its low power and 1240kg mass. The electric motor’s low-speed torque and petrol engine integrate seamlessly, giving punchy performance in all but one driving situation. That situation is climbing a long, steep country road hill. Initial performance is fine but after a while, the high voltage battery becomes drained, decreasing the amount of electric power available. The car slows, the engine revs automatically increase, and you can find yourself with the accelerator pedal flat to the floor, just crawling along. At this point a rather cute tortoise symbol lights up on the dash – we christened her ‘Myrtle the Turtle’. And it just so happens that I live at the top of a long, country road hill that has gradients as steep as 16% – and Myrtle is a real pest. By feathering the throttle and watching the movement of current in and out of the HV battery (the colour LCD shows this), it was possible to negotiate this hill with a speed over the crest of 55km/h or so. But lose concentration and that dropped to 47km/h. That’s awfully slow. (It should be stressed that the later model Prius cars don’t suffer from this hill-climbing problem.) I considered adding another HV battery pack but because of their control systems, this is an extremely complex move – nothing like as simple as wiring the two batteries in parallel. That left increasing engine power as the best option and so I turbocharged and intercooled the engine. And then the fun started…. Fascinatingly, the hybrid control system coped with the increased power remarkably well. Presumably because Toyota’s engineers could never predict exactly how much power the petrol engine would generate (as with all engines, this varies with atmospheric conditions, individual engine build quality and so on), the hybrid system had enough flexibility in its control system to direct any excess power being developed by the engine into charging the HV battery. Note: into the HV battery, not to the wheels… So the odd situation developed where, with the HV battery level at (say) half, the performance of the car was standard – despite the turbo! But what about up the big hill? Ah, well there the car was transformed. Rather than the HV battery dropping in charge, going up the hill it actually From the publishers of SILICON CHIP PERFORMANCE ELECTRONICS FOR CARS NOT A REPRINT: More than 160 pages of new and exciting projects never published before – all designed to get top performance from your car. FASCINATING ARTICLES: 7 chapters explaining your car – engine management, car electronics systems, etc ADVANCED PROJECTS: You’ll build controllers for turbo boost, nitrous, fuel injection and much more! We explain the why as well as the how to! Available direct from the Publisher ($22.50 inc postage): Silicon Chip Publications, PO Box 139, Collaroy NSW 2097. Ph (02) 9939 3295; Fax (02) 9939 2648; email silchip<at>siliconchip.com.au or via our website: www.siliconchip.com.au siliconchip.com.au March 2006  19 A thick-walled exhaust manifold was fabricated to locate and direct exhaust flow to the added turbocharger. The mechanical addition of the turbocharger was relatively straightforward – it was getting the air/fuel ratio suitable for the forced aspiration that caused Julian Edgar to pull out his hair…. increased in charge! And always having lots of battery voltage – and so electric motor power – resulted in the speed over the crest of the hill increasing from a low of 47km/h to a stunning 86km/h. But let’s go back a little. When the turbo was fitted, the air/fuel ratios needed to be changed – a turbocharged engine (especially one with a 13.5: 1 static compression ratio, Atkinson or no Atkinson) needs richer than standard mixtures. The Prius uses an air/fuel ratio of 14.7:1 all the time – the engine ECU monitors the output of the two oxygen sensors to constantly hold this air/fuel ratio. So how to change this? The first step was to fit a Digital Fuel Adjuster (from “SILICON CHIP’s Performance Electronics for Cars”) to allow alteration of the airflow meter output. However, monitoring the mixtures with a MoTeC air/ fuel ratio meter showed that the Prius ECU is extraordi- The Prius airflow meter sensing element is normally built into the airbox. With the addition of a turbo, the standard airbox could no longer be fitted and so the airflow meter sensing assembly was removed and built into this new, larger airflow meter. The SILICON CHIP Digital Fuel Adjuster (DFA) was then used to correct the output signal. 20  Silicon Chip To provide the additional fuel required for the turbo, a dual-pressure fuel system was installed. This uses two pumps, two pressure regulators and a solenoid to switch from low to high fuel pressure. The switching is done by a SILICON CHIP Simple Voltage Switch that monitors airflow meter output voltage. When load rises sufficiently, the fuel pressure is switched high and at the same time, the twin oxygen sensors are disconnected. This approach prevents the engine Electronic Control Unit ‘learning around’ the changed mixtures. narily quick at learning around any changes made in this way. Alter the mixtures to 12.5:1 and within five or so seconds, the mixtures are back at 14.7:1! The same thing occurs if fuel pressure is increased – back go the mixtures to stoichiometric. Hmmm, so what about disconnecting the oxy sensors, substituting an appropriate looking 0-1V square wave signal on the ECU oxy sensor inputs and then altering the airflow meter output voltage? A pair of 555-based circuits was constructed and the system wired-up. But the ECU immediately picked that something was wrong with the oxy sensors and went into a default mode – which with the added airflow of the turbo, resulted in mixtures even leaner than 14.7:1! Aaaaagh. Well then, what about disconnecting the oxy sensors (easily achieved just on high load with the “Simple Voltage Switch” [also from SILICON CHIP Performance Electronics This 555-based circuit was built in an attempt to simulate the fluctuations in oxygen sensor voltage normally seen by the Electronic Control Unit (ECU). If successful, this would allow the oxy sensors to be disconnected, stopping the ECU learning around mixture changes. However, the ECU immediately saw through the pretence, outputting the same mixtures as when there is no oxy sensor input at all. siliconchip.com.au These logged traces compare the input and output of the brake pressure modification circuit. As can be seen, the output trace shows a lot more ‘area under the curve’, indicative of a higher brake pressure being monitored by the Electronic Control Unit. The result is increased regenerative braking which improves economy and reduces friction brake wear. for Cars] working off the airflow meter output voltage) and then using the Digital Fuel Adjuster (DFA) to alter the mixtures? Again, no success – and this time, the mixtures appeared to vary randomly. About this time I upgraded the fuel system with a new in-tank pump, external adjustable pressure regulator and a return-line to the tank. That allowed me to run higher fuel pressure (which initially gave correct mixtures with the oxy sensors disconnected) with the DFA used to tweak the resulting mixtures. But, yet again, the mixtures were not consistent. After many weeks of work, I finally devised an effective system. Two fuel pumps and two fuel pressure regulators are used to allow the running of two different fuel pressures. A solenoid allows electronic switching between the two different pressures. The lower of the two pressures is set so that, even when running a little turbo boost, the ECU can keep the mixtures at 14.7:1 and the oxy sensor feedback loop operates as normal. Then, when a preset load is reached, a Simple Voltage Switch monitoring the airflow meter voltage switches out the oxy sensors and activates the solenoid increase in fuel pressure. The resulting mixtures are then fine-tuned by a DFA working on the airflow meter output. With this approach, the mixtures are consistent and economical (the car is still in closed loop with 14.7:1 mixtures for the vast majority of the time) but with appropriately rich mixtures used at full turbo boost. (Incidentally, ignition timing has never been a problem. I have never heard the engine detonate, even when [briefly!] running 15 psi boost. 95 octane fuel is used – the car is designed to run on 91.) Auto throttle shutdown About this stage I started to relax. Ahhh, this is nice...the world’s only turbocharged, intercooled Prius with modified regen braking. Those many hours of work were well worth it. And that was the case until I discovered that the hybrid control system’s adoption of turbo power wasn’t as seamless as I’d first thought. Initially, I’d decided the abrupt engine shut-down that occasionally occurred at full power was an ignition problem – and had replaced the spark plugs with a colder heat-range Iridium design. siliconchip.com.au March 2006  21 But then, while watching the boost gauge, I saw what was happening. At full power, the hybrid system would momentarily close the electronic throttle. Whether that’s to protect one or both of the electric motor/generators – or for some other reason – I still don’t know. But the result was a huge power loss perfectly timed to occur when overtaking a semi-trailer…. If the shutdown was a result of excessive engine power, I could just drop turbo boost at higher revs. But the problem was that I couldn’t. The waste gate spring pressure on the turbo meant that 7 psi was as low as I could go. But there had to be another way of dropping boost – and there was. By using a solenoid to control the boost pressure feed to a recirculating blow-off valve, the valve could be made to leak, bleeding boost from the compressor outlet back to the inlet. The result is decreased boost. I initially used the Performance Electronics for Cars Independent Electronic Boost Control (IEBC) to achieve this function but ran into a snag. The IEBC allows the mapping of a pulse-width controlled solenoid on the basis of engine load, as measured by injector duty cycle. However, it doesn’t have an adjustable hysteresis function and so the solenoid would close (causing the blow-off valve to open) and then as a result of the lower boost, engine load would drop, thus switching the solenoid back on! Boost would then surge up and down. (This isn’t normally a problem with the IEBC because the slope of the adjustment curve can be made gentle. But in the case of the Prius, it had to be a much more sudden change.) The answer was to (again!) use that most ubiquitous of building blocks, the Simple Voltage Switch. With its adjustable hysteresis, it was ideal in this application, wired-in to monitor airflow meter output signal. As finally configured, boost rises to 7 psi and then above the preset load point, smoothly drops to 4 psi. This makes no difference to on-road performance – and there are no throttle shut-downs! Auto engine off While compared with many turbo applications the Prius turbo is not working particularly hard, there is one characteristic of its hybrid control that if left unaddressed, could quickly kill the turbo bearing. As described earlier, when the throttle is lifted, fuel flow to the engine is stopped. So, approaching a red traffic light, the engine stops running as soon as you back off – and stays off until the lights go green and you apply the accelerator. A turbocharger relies on engine oil flow for lubrication and partly for cooling, and so the engine should not be turned off until the turbo has had time to cool. If the turbo has been spinning hard, an early engine switch-off can cause oil in the turbo bearing to coke. So how could the Prius petrol engine be kept running after a boost event? This model Prius has two air-conditioning modes. In High mode, the engine is forced to run continuously. In Normal mode, the engine is allowed to switch off whenever the hybrid ECU decides it should be off. When High mode is selected, the air-con system tells the hybrid ECU that it should not switch off the engine by means of an ‘engine on’ request signal. This signal is very simple – above 4V means keep the engine running, below 1V means it’s OK to switch it off. 22  Silicon Chip To prevent the hybrid control system automatically switching off the petrol engine shortly after the turbo has been on boost – and so starving it of oil – a Simple Voltage Switch was modified to provide the ‘engine on’ request signal than normally occurs when the air-conditioner is switched to high. This photograph shows the new MOSFET, capacitor, diode and resistor that allows the Simple Voltage Switch to feed the correct signal to the hybrid ECU for 30 seconds after the turbo has been boosting. So by feeding 5V to the ‘engine on’ input of the hybrid ECU, the engine can be kept running. Yet another Simple Voltage Switch was used to monitor the airflow meter signal, sensing when engine load was above a certain threshold that corresponded to a few psi of boost. To derive the required 5V signal to feed to the hybrid ECU, a pot was placed across the regulated 8V supply used in the Simple Voltage Switch. The Simple Voltage Switch was then modified so that once it was triggered, it stayed on for about 30 seconds. So how well does the system work? In most cases the delayed ‘on’ time for engine running isn’t noticeable – when the car is moving, it’s hard to tell whether the engine is firing or not. But if the car is being driven hard around the city, pulling-up at a red traffic light causes the engine to keep running for a short time, when previously it would have switched off as soon as the car was slowing for the stop. Conclusion We’ve run out of space to cover all the mods made to the car – they include a rear sway bar, electronically modified electric power steering, on-dash mixture indication, electronically controlled high pressure intercooler water spray, under-floor aerodynamic changes and plenty of other bits and pieces. And the results? Well, the worst aspect of the car was previously its country road hill-climbing ability – and that’s been improved by over 80%! But what about the raison d’etre of the Prius: fuel economy? The modified Prius now has better than standard fuel economy. On an open road cruise at 100km/h, the turbo Prius will turn in a best economy of about 5.5 litres/100km, whereas in the same conditions, the dead standard car used to get about 6.3 litres/100 km. And even when being driven hard, the economy now averages about 5.8 litres/100km. SC siliconchip.com.au