Quote:
Originally Posted by clarence
There are 2 versions of the U0 engine (U0 is 20i's state of tune whereas O0 is 28i's state of tune) - high compression version & low compression version. The HC version has comp ratio of 11:1 whereas the LC version has comp ratio of 10:1 (i.e. same as O0). As confirmed by BMW China tech staff, the HC version have different piston crowns. They also confirmed that detonation will occur if the HC engine is ran at O0's (i.e. 28i's) level of boost.
Hence tuners cannot get O0's level of output using this version of U0 with just a piggy-back (AFAIK the DME on the N20 has not been cracked). Major upgrades need to be done to reach around 235ps (confirmed by poster several pages back).
The LC version of the U0 can theoretically be made to output at O0's level of output by running at O0's level of boost (or close to it's rated output if binned due to some "defect"), but it seems no tuner have tried to distinguish between these 2 versions of U0. There can be several reasons for this, e.g. this engine is only fitted to E84 & F25 during certain production dates, & for certain production dates it also had a different turbo, so it may be difficult to find a development car. As BMW themselves don't disclose why each of these LC engines are classified as U0 rather than O0, the tuners had to play safe. Again as the DME has not been cracked so all they can offer is a piggy-back.
The testing u mentioned is already done during prototyping stage, wht they do at the plant is a dyno run. A failure can be caused by sth as simple as the engine missing a few Nm's compared with the rated torque at idling speed.
|
Thanks for the summary, I was following the thread quite closely, but it's nice to have all the info in one thread.
But let me rephrase my criticism against your binning hypothesis in a clearer form. I am not rejecting the notion that there would be some binning. I am arguing that that alone is not enough to explain why the low output version cannot be made to generate higher output for long enough for someone to develop and sell the software.
In other words, if you apply a cut on a continuous distribution with one maximum, most of the failed cases will come from just below the cut line, meaning that they will be very similar to the ones that barely passed the cut. Now, we know that the ones that pass the cut is very likely capable of withstanding high output for 100K miles or longer. Then there is no way that the ones that barely failed the cut, which is the most likely case, will be so different from the ones that barely passed, that no tuner can tune it up long enough to market a product.