and here's the basis of our argument over air-bleed "fuel-saving" devices.

there are basically two reasons.

1. the device is doing no good since additional fuel is being injected to compensate for the addt'l air

2. the device is saving some fuel and reduces HC & CO levels - but dramatically increases the output of NOx.

it doesn't take an engineer to absord all this stuff.

[quote]

The ratio of air to fuel is a critical factor in petrol combustion. The theoretically ideal mixture is around 14.5 parts of air to 1 of fuel. This is known as the stoichiometric ratio. Variations from this are characterised by the term lambda, where lambda less than 1 means a rich micture (excess fuel), and lambda greater than 1 means a lean mixture (excess air).
The first reason is the presence of the lambda sensor in the exhaust system. For good operation of the catalyst (see below), the engine is set to run at lambda 1 at all times other than cold start and full load (in other words, for 99% of the engine's operation). Since the tolerance on injectors and air meter might be as much as 5%, a lambda sensor is fitted in the exhaust to monitor the mixture. This sensor gives a very precise measure of lambda; the engine management system adds or removes a small amount of fuel to keep the engine at exactly lambda 1. So the air bleed through the "fuel saving device" results in a lean mixture, but the lambda sensor will "see" this and the EMS will put extra fuel in to compensate. Hence the mixture stays at lambda 1 - the "fuel saving device" has no effect at all

The second reason is the operation of the catalytic converter. This miraculous device is largely responsible for the significant improvements in urban air quality seen in recent years, by converting around 95% of toxic pollutants to relatively harmless substances.
The three "regulated pollutants" are unburnt hydrocarbons (HC), carbon monoxide (CO) and oxides of nitrogen (NOx). All three are serious human health hazards. NOx is nitrogen with added oxygen; HC and CO represent incomplete combustion. What the catalyst does is to remove the oxygen from the NOx and give it to the HC and CO. The result is plain nitrogen, carbon dioxide and water, all of which are essentially harmless. But to work, the lambda value must be just right. If the mixture is too rich then the oxygen in the NOx is not enough to give to the HC and CO, so they pass through unconverted. If the mixture is too lean then the "spare" oxygen is given to the HC and CO, and the NOx passes through unconverted. The overall effect of lambda on post-catalyst emissions is like this:



There is a very narrow "window" between lambda 0.99 and lambda 1.00 where all three pollutants are effectively cleaned up. It is precisely because the engine must operate in this "window" that modern engines are fitted with lambda sensors.
Now, under most circumstances the lambda sensor will correct for any fuelling deviations introduced by the "fuel saving device". But the sensor has limited "authority" - it can only add up to about 10 - 15% additional fuel, since otherwise a faulty sensor could result in an extremely rich mixture, leading to engine problems. If the deviation exceeds this limit, then the lambda sensor can no longer compensate and a lean mixture is the result. So the result of fitting one of these devices may be a lambda around 1.05 - 1.10. This will give a small improvement in fuel consumption (about 2 - 5%) and, as the graph above shows, reductions in HC and CO. This sort of result is often shown by makers of such devices as "proof" of emissions reduction. However, the NOx emissions will increase massively - perhaps 20 times - as the catalyst can no longer clean them up. Since NOx is not usually measured by workshop test equipment, this is not picked up. The overall effect is a very large increase in toxic pollutants, which does not fit the "green" claims of such devices.