The Story of Turbo'ing a Hybrid Prius, Part 1

In electronics it’s called the Smoke Test. Circuit wired-up and complete, the reality check comes when you switch it on.

In turbo fitting it can be called the Boost Test. As in, how much will be developed where in the rev range? If it’s a known engine/turbo combination, the Boost Test is mostly a case of proving the absence of exhaust or intake leaks. But when it’s a completely unique matching of a turbo to an engine, the test assumes critical importance. Will this system work at all?

So it was with more than usual attention I watched the boost gauge, weeks of work hanging in balance....

The forced aspiration project carried out on the car - a ’99 Toyota Prius hybrid – had so far been a bizarre mix of success and failure.

The first step had been to fit a positive displacement supercharger, one of the tiny ones normally fitted to a Subaru Vivio (or was it from a Rex?). Getting the blower under the short, stubby bonnet of the Prius had become a near-nightmare of precision engineering – the clearances being so tight that the supercharger bracket had to do double duty as an engine mount. The intake plumbing to the blower – made from 2-inch copper tube – twisted and weaved its way past an intake manifold bolt and the variable valve timing solenoid; the elbow ended up with so many depressions in it that flow must have been harmed.

However, once the blower was in place, I was very pleased with the job – the belt ran perfectly, the mount was rigid and relatively light in weight, and the bonnet still closed over the intercooler plumbing. Just.

And boost? Well, it was available from idle to redline, the latter being 4000 rpm in this 1.5-litre engine. By adjusting the recirc valve, I could have it anywhere from 6 or 7 psi downwards; in the final iteration I ran about 6 psi.

And how did the immensely complex battery/electric hybrid control system cope with air being shoved through the electronic throttle? ‘Seamlessly’ is the best way to describe it. Rather than there being increased top-end power, the hybrid system simply pumped more charge into the high voltage storage battery, resulting in far more average electric power being available to drive the wheels. With the electric motor being nearly 70 per cent as powerful as the petrol engine, having this power always on tap gave the desired performance outcome.

So where was the failure? It can be summarised in one word: noise!

The supercharger whined like a police siren. Initially it even did so at idle, but I managed to quieten the intake noise with a more restrictive airbox and by juggling the recirc valve opening. But even then, when you put your boot into it, the siren under the bonnet was loud enough to turn people’s heads 20 metres away. I tried every trick in the book to quieten this pressure pulsing noise, but nothing brought the noise level down to what I considered acceptable.

In the end I made the agonising decision: take off the blower and instead fit a turbo.

Turbo Selection

But what size turbo to fit? As covered in more detail at Driving Emotion, the Prius engine is unique, with this model running a 43kW 1.5-litre engine. The small peak power output is primarily because of the low redline; however, it’s also an Atkinson cycle design, a valve timing approach which affects the mass of air that gets breathed each intake stroke.

After considering all the options, I went for a secondhand turbo – one of the two from an EJ20 Subaru B4 engine. In standard form, the Subaru 2-litre four-cylinder engine develops about 190kW. From that it can be assumed that each turbo flows about 95kW; that is, each turbo can be thought of as being from a 2-litre, 95kW, 3500 rpm redline engine that uses about 12 psi boost.

That’s more boost than I need in the Prius application (I was quite happy with the results gained from the supercharged 6 psi), so the question became: would a turbo that produces 12 psi boost (and is fast to boost) on a 2-litre, 95kW, 3500 rpm redline engine be suitable for a 1.5-litre engine with 43kW and a 4000 rpm redline? The immediate answer is ‘no’, that the much smaller turbo from a 660cc Kei class Japanese car would be better - those engines have a peak power of only 47kW. But it gets more complex: in the Prius application, I needed only half the peak boost pressure of the Subaru...

The single turbo from the EJ20 twin turbo is also very close in size to the pictured Garrett GT12 – a turbo that on paper looks fine for the Prius. And with the Prius’ transmission – which acts very much like a CVT – the engine revs can flare quite high even at low throttle angles. That would help the turbo spool up, but then again there would likely be quite a lot less exhaust gas available to do the work.

So the ex-Subaru turbo was it – but at best its selection was only an educated guess. AUD$150 later, I had one in my hands.

Turbo Modification

The IHI RHF4 turbo might look like it was sized correctly, but in physical layout it was all wrong for the Prius.

In the pictured Subaru application, the exhaust feeds comes up from below, whereas in the Prius the exhaust manifold would be coming in from above. This turbo uses a V-band to connect the turbine to the centre housing, allowing this end of the turbo to be rotated. And since the compressor housing is bolted into place, it seems easy enough to undo everything and spin it all around until all the entrances and exits are in the right places. Unfortunately, though, the reality is different. Yes, the turbine can be rotated, but this in turn places the mounts for the wastegate actuator in the wrong place. Yes, the compressor cover can be unbolted, but it cannot then be rotated – the compressor bolts into place in only the one orientation!

In the end the turbine housing was rotated, which necessitated a new mounting arrangement of the wastegate actuator - it normally bolts to cast-in bosses on the compressor cover. To provide a new mounting surface, my newly-acquired lathe was used to turn-up a thick aluminium ring. This was sized to be a snug fit over the compressor cover, with a section of the ring cut out to allow the compressor discharge nozzle to fit through it. The ring was held in place by longer-than-standard stainless steel cap bolts, the ones that also hold the compressor cover in place. Spacers were used to distance the ring from the compressor lugs – without these, the ring would have had to be a very complex shape to provide sufficient ‘meat’ around the bolt-holes. The wastegate actuator was then bolted to the ring.

Getting all this right was an unexpectedly complex procedure – with a commensurate expenditure in time. But when it was finished, the wastegate moved smoothly in its new orientation.

Turbo Exhaust Manifold

The turbo exhaust manifold – the supposed stumbling block of custom fitting a turbo – proved to be almost ridiculously simple to make. Well, simple in comparison to the supercharger mounting bracket, anyway...

The first step was to cut out a flange from 12mm plate. Using the exhaust manifold gasket as a guide, this was done by Peter, the local welder. (By this time, Peter knew me pretty well – he’d done the cutting and welding of the supercharger bracket, and the brazing together of the intake and intercooler plumbing.) The next step was to piece together the runners, which were made from 1¼ inch ID buttweld fittings. With only a few interim welds along the way, this was easily achieved using a hacksaw and a round file to shape the bends so that they nestled together. After the final welding was done, the welds ground back, and the manifold sandblasted and then faced, I was pretty pleased with the result.

I emailed a photo of the manifold to a friend, Paul. But his reaction had me aghast. After some initially admiring comments, he ICQ’d: I *thought* the way to design the manifold would have been based on the firing order and path length. Will there be a conflict between [cylinders] 1 & 3? The manifold was finished, but here was someone suggesting it wouldn’t work...

I’d employed four criteria when laying out the design: (1) it had to position the turbo in the right place, (2) the manifold had to easily fit in the available space (the turbo is positioned between the engine and the firewall), (3) the runner diameter was to match the port diameter, (4) the pipe junctions had to ‘flow’ together. I had not specifically considered the firing order of the engine, and nor had I considered that long, equal-length runners were a priority. The latter contrasts with most aftermarket fabricated manifolds, where people place a huge premium on having long runners of equal length. However, I am always much more impressed by factory engineering and almost no factory turbo exhaust manifolds are made in this way. And neither are the aftermarket cast manifolds developed for mass-produced turbo kits.

I looked in all my turbo reference books on the subject of turbo manifold design, I looked at engineering papers, and I looked at all the factory turbo manifolds I could find. I also looked at the factory Prius exhaust manifold, which in layout and pipe size is very similar to the manifold I had made. And in the end I decided that it would work fine... but a doubt had been planted. Not only might the turbo be too big, but now the manifold might not be effective...

Next week: will the turbo system work?