When lower boost doesn’t slow a car…

Posted on December 11th, 2005 in Opinion by Julian Edgar

As regular readers will know, my NHW10 hybrid Toyota Prius has been turbo’d and intercooled. To fit in the available space, the turbo – one of the puffers from a twin turbo Subaru Liberty – required that its wastegate mounting system be modified. A spacer ring was used to allow the wastegate actuator to be placed in a different orientation to standard. This ring put a small preload on the wastegate rod, resulting in a minimum on-load boost level of 7 psi. That’s 7 psi, even with the wastegate hose connected directly to the turbo compressor outlet – ie no bleed or aftermarket boost control fitted.

Now that’s generally well and good, but sometimes at high loads, 7 psi can cause a problem. Intermittently – and for only a very short duration – the hybrid control electronics closes the electronic throttle. I assume that this occurs because I have exceeded a preset internal safety trip-point for the engine or electric motors. That implies that if boost can be dropped a little at the top end, the throttle shut-downs are likely to stop. (And in previous short-term testing with lower boost levels, the problem did in fact disappear.)

Since the minimum boost level the wastegate can be set to is 7 psi, dropping boost below that requires bleeding air from the manifold. One easy way of achieving this is to allow the blow-off valve to leak, something which can be achieved by pulse-width modulating the boost/vacuum feed to the valve. Working with the airflow meter signal, the Simple Voltage Switch kit allows this boost drop to be triggered at a preset load. This boost leak doesn’t cause any fuelling problems, because a recirculating blow-off valve is used and the air is returned to the intake after the airflow meter.

Using this approach, I initially dropped peak boost back to 5 psi at loads over about 80 per cent of max, with little discernible difference in performance. Since dropping from 7 psi to 5 psi apparently made little variation in the available top-end power, I then decided as an experiment to allow the blow-off valve to leak all the time. (The aftermarket GFB valve runs a variable preload on its internal spring, making this easy to achieve with some spring adjustment and pulling off the vacuum/boost feed hose.) This resulted in a slow rise in boost to a max of only 4 psi.

Of course, slow rising boost is an anathema in a turbo car – you always want boost to come up as fast as possible. In a normal car, the difference between this and the previous boost-as-fast-as-possible-to-7-psi-and-then-hold-it-at-that-level would be like chalk and cheese. The configuration with the slow-rise-to-4-psi would feel half-dead and power would be clearly way down.

For Godsake, for some testing forget the bloody dyno – get out on the road!

Posted on December 4th, 2005 in Opinion by Julian Edgar

Why is it that people put so much faith in dyno testing? I have written about this topic before (see Driving Emotion – August 2004) but it needs to be continually shouted from the rooftops. Dynos are bloody useless in so many areas of car modification testing that I don’t know even where to start. But I’ll try.

As I wrote in that previous column, they’re pretty well useless for testing turbo boost controls. Why? Well they:

  • Don’t take into account the acceleration rate of each gear – vital because boost overshoot on transients is hugely affected by the rate of engine rpm increase.

  • They don’t allow the testing of boost behaviour of full-throttle gearchanges (very few people do full throttle gearshifts on the dyno). Again, it’s in just these conditions that you look for boost overshoots and/or slow increases back to peak boost after each gearchange.

  • No one ever does a full-bore launch from a standstill on a dyno. And the speed with which boost can be brought up in these conditions – ie controlling wastegate creep – is a major aspect of good boost control.