The most widely-quoted test data for the Ecotek CB-26P is the report from the Warren Spring laboratory (formerly part of the DTI). This was conducted in 1993 and relates to a carburetted vehicle from 1988 with over 110 000 miles on the clock. As is made clear in the main Case Study, benefits to a vehicle of this kind do not in any way suggest benefits to a more modern car.

Even neglecting this, the Warren Spring test is fundamentally flawed. It is entirely accepted by both industry and governments throughout the world that any meaningful emissions and economy measurements must be carried out in a rolling-road emissions laboratory, where standard driving conditions can be maintained. The Warren Spring test data was collected by driving a vehicle round in a busy town - every motorist knows that traffic conditions vary from day to day and indeed hour to hour, so the driving conditions of the test car would be entirely different on each test and hence it is impossible to say if the improved economy is down to the device or simply varying road conditions.

To be fair to Warren Spring, they did recognise this problem and devised a test strategy to obtain the most meaningful results possible under the circumstances. The device was tested on two cars, which were driven round the test circuit together for each run and so were subject to approximately the same traffic conditions (hence the reference to "leading" and "following" in the data shown on Ecotek's website). Put crudely, the device was fitted to one car for four tests, and then removed and fitted to the other car for the next four. If both cars showed an improvement with the device fitted, this could be said with some confidence to be due to the device. If one showed an improvement and the other a worsening, this would probably be down to changing traffic conditions.

While one car (for which Ecotek present test data on their web site) showed an improvement, the other car had significantly worse economy, HC and NOx emissions when fitted with the device. This fact, which Ecotek conveniently neglect to mention on their site, strongly implies that the benefits seen on the first car were largely due to changing test conditions. The full test data is as follows (bold indicates device fitted):

Run	Blue car consumption	Red car consumption
1	13.07	15.04
2	8.97	11.76
3	10.61	11.44
4	9.54	14.98
5	8.76	9.90
6	8.07	9.90
7	9.41	8.07
8	10.43	11.22
9	8.57	11.96
10	9.39	11.59
11	9.64	11.40
12	12.30	12.48
13	9.81	12.16
14	9.90	12.19

This doesn't really show very much, but the data can be grouped into sequences to make it clearer. It is also useful to add in the average speed for each sequence, which is given in the full Warren Spring report:

Runs	Blue car consumption	Red car consumption	Average speed (kph)
1-4	10.52	13.31	19.7
5-8	9.17	9.77	25.9
9-12	9.97	11.86	22.1
13-14	9.85	12.18	22.6

Since this is a city drive, a higher average speed suggests less time spent stationary (at idle), and therefore a better fuel economy could be expected. Plotting economy against average speed for each car does indeed show a very clear trend:

Especially interesting is to look at the circled data points, which represent the tests with the CB-26P fitted. They lie almost exactly on the trendlines - which suggests that the device has no effect on fuel economy at all! It just so happens that the runs where the device was fitted to the red car took place when traffic was light, and so the results were better than average. By contrast, the runs where the device was fitted to the blue car took place when traffic was heavy and so the results were worse than average. Unsurprisingly, where fuel consumption was lower (and indeed the total test time was lower), all the toxic emissions are also reduced.

The data for the blue car is not wholly reliable, since its ignition timing was apparently adjusted at the same time as the device was fitted. However, if anything this could be expected to deliver even better economy benefits, since the ignition timing was presumably adjusted to the optimum setting having previously been incorrect. The fact that all the results lie on such a precise trendline also suggests no significant change in the vehicle.

Of course, all this may just be a coincidence. But given that the vehicles were very similar in terms of their engines and fuelling technology, it would seem likely that the device (if it truly worked) would have the same effect effect on both cars, rather than making one better and the other worse. Certainly the Warren Spring test is far less conclusive than would first appear, when the full report is analysed in detail.

This analysis is of course especially relevant to the Ecotek CB-26P, but also serves as an illustration as to why on-road test data is not a good way to assess fuel "saving" devices. Even when a professional research body carefully tries to keep to controlled test conditions, natural variability is so high that any possible effect of the device on test is entirely masked. The same problem applies, only even more so, to tests conducted by journalists or buyers of such devices.

Another data set widely reproduced on the Internet is the fuel consumption figures from Evans Halshaw. These tests consisted of measurements of fuel consumption at steady speed in top gear over a mile on two cars, a Vauxhall Carlton 1.8 and Vauxhall Cavalier. MPG figures are quoted with and without the Ecotek valve fitted, but look at the "baseline" (unmodified car) data:

Cavalier: 25.5 mpg @ 50 mph, 20.2 mpg @ 70 mph
Carlton: 11.5 mpg @ 40 mph, 23.8 mpg @ 60 mph, 24.3 mpg @ 80 mph

(data for the Carlton was removed from Ecotek's website in February 2005, but can still be easily found on the Web, for example here)

Two things are immediately noticeable in these figures. Firstly, the Carlton apparently gets more economical at higher speeds, which is the complete opposite of what would be expected. Cars are at their most economical at a steady speed of around 35 - 50 mph (perhaps slightly higher for automatics) in top gear and get steadily less economical at higher speeds. This is mostly because wind resistance increases by a factor of four when the speed is doubled, and is true of the vast majority of vehicles. Hence the data presented by Ecotek, showing the opposite trend, is very suspect.

Secondly, the figures overall are extremely poor. "Constant speed" fuel economy figures are not officially measured any more, but looking up some old data gives these figures:

1.6 Mondeo (similar to Cavalier): 51 mpg @ 56 mph, 39 mpg @ 75 mph
2.0 Omega (similar to Carlton): 43 mpg @ 56 mph, 34 mpg @ 75 mph

These "official" figures are also not far off what one would expect on a steady motorway run and so are quite believeable. Yet the "Ecotek" cars - apparently at a steady speed on a flat road - used around twice as much fuel. In the case of the Carlton, it apparently achieved an astonishingly bad eleven miles to the gallon at a steady 40 mph in top gear. So either the cars were in very bad condition to start with, or there was something wrong with the test method. In either case the improvements with the Ecotek fitted are therefore of little significance. (This apparently faulty data, by the way, appears to be the main basis for the "30% economy gain at high speed" and "gets better the faster you go" claims that are often seen on the Net.)

Ecotek themselves, to their credit, no longer cite these tests (Warren Spring or Evans Halshaw) as proof that their device works - nonetheless you will often see the data on the Internet and it is useful to understand its flaws.

Go back to general comments about the Ecotek CB-26P