Thanks for your patience Shopp, same excuses for late reply on your last post : loads of work and a family to care for...
Finally had a chance for more pipes cleaning and some more tests. In short, it's still not solved but there's clearly some progress.
1/ As my MAFs connectors were in bad shape, I had them rewired but did ask for a 10cm extension on each side. This way I'm now able to swap the connectors between MAFs. The verdict is clear : it's definitely a mechanical problem, nothing to do with the wiring, ecu or whatever ground / resistance issue.
2/ Following my last pipe cleaning campaign, the fault code related to the MAFs discrepancies has not reappeared yet, which is obviously great news indeed!! More precisely the MAFs reading at idle are still not good (20-40% difference) but under load they seem to behave better. I'll come back with more precise measurements to substantiate this.
3/ Fuel trims, lambdas are still perfectly aligned (they always were, but glad it's still the same)
4/ Cats on both side have the same temperature so the clogged cat possibility should be ruled out
In addition to the above points, my mapper suggested a few routes to follow :
a/ swap dump valves as they may not be properly calibrated (actually I've changed them for new oems, but didn't swap them within a pair). Although the probability is low that this could be the source of the problem, it's quite easy to try
b/ test compressions : already discussed with Shoppinit. If compression on a cylinder was that bad to allow for such discrepancies, the engine would run very badly. Nevertheless it's always good to evaluate the sealing.
c/ discrepancies between turbos : after 160,000km of use, turbos may not perform equally and according to my mapper this could well explain the difference between MAFs. How does that sound to you?