Before the turbo in the 5 branch collector.

Surely, the ideal position would be a seperate sensor in each branch of the manifold, so that you can see variances in each cylinder's fuelling

Flea - I learn something every day - always thought a cat would screw the readings for a wideband

If we are talking ideal then you should have a wideband and EGT at the exhaust port of each cylinder. Each cylinder is effectively a separate engine.

The cat has to do something to do with AFR readings otherwise modern cars wouldn’t have a lambda before and after unless the second sensor isn’t measuring O2 but NOX? I have never played with a car with a CAT so can you enlighten me?

A lambda sensor only measures oxygen no NOX, it compares the oxygen level in the exhaust with the outside air.
It seems to me that it's possible that a cat does effect the oxygen level in exhaustgas. Because of the reactions taking place.
oxidation
2HC+2O2--> H2O+2CO2
2CO+O2--> 2CO2
reduction
2NO-->N2+O2

Well most of the cars I map have catalytic converters and of course I invariably use a wideband in the tailpipe. They will all quite happily idle and cruise at stoich exactly the same as the narrowband.

I know a lambda sensor only measures 02 but Fleas comment says that a CAT makes no difference to lambda readings so I was asking that if the second sensor is a NOX sensor instead of an 02, or enlighten me as two why manufactures put a lambda before and after the cat if it makes no difference to AFR’s?

That seems logical to me because as you say at idle and cruise you have a stoich mixture, which means a complete burn of fuel. At this moment your exhaust gas consists of H2O and CO2, and the cat doesn't participate.

But when your running rich your exhaust gas also contains HC (unburnt fuel) and CO. At this point the cat comes into the equation, meaning the following reactions which use up oxygen.
2HC+2O2--> H2O+2CO2
2CO+O2--> 2CO2
Which in effect means your exhaust gas after the cat contains less oxygen than before the cat, and since a lambda sensor measures the oxygen level this will result in a measurement that reads a bit richer than before the cat.

Offcourse this is all theory, and I don't know how this stacks up to reality. It could be an interesting experiment to do a run with one wideband before the cat and one after the cat at he same time and compare the readings.

The second sensor is still an O2 sensor, it's there to check if the cat is working correctly. The value of the second sensor shouldn't vary to much if the cat is working correctly.

Don't forget an MOT test checks air fuel ratio or lambda, which needs to be at stoich with or without a catalytic converter. However, even at WOT the cat doesn't affect readings.

Wideband manufactures seem to think the cat effects the reading.
Bottom of page 4

FWIW, my coop has a wideband stack, just before the CAT & Flea has mapped it using a wideband in the ass end, basically, they are pretty much the same near stoich, however there is a small discrepancy on boost, but again, this DEPENDS on how HOT the cat is, if it's cold, no USE.

On my last MOT, it passed emissons after Flea mapped, even on fast idle with a 2.4 & normal go-go juice.

Interesting Taz, I'm not trying to discredit Flea, in fact I think he does a incredible job considering his results. I'm just intrigued by the subject.

TurboJ, here is a small article why manufacturers run two lambdas.
http://www.fordscorpio.co.uk/catmonitor.htm

No probs pal, it was just for info