Since the 250 GB chipset is close to being available, more features, less bugs, with PCI-lock, ..., why considering the 150 chipset now?
Plenty of nF3 150 boards have PCI locks. I'm listening to music right here at what would be a 50MHz PCI bus on my SB Audigy2. The only bugs I've encountered are specific to DTR laptop processors, AFAIK, because no one else has, and even these are insignificant at best.
So, we've found two "features" that the 250 will offer over the 150, the first being the 800MHz HT, the second being a slightly improved SATA RAID controller. The nF3 150 can usually handle HT speeds into the 700MHz range anyways, and the performance difference is minimal. I don't use four hard drives, and I doubt most others do either. For me, RAID 0+1 is a moot point; I can't afford four HD's anyways. If one must use four hard drives, the PATA RAID controller supports RAID 0+1, and there is a very minimal difference between SATA and PATA performance. For these reasons, I believe that few will be affected by the very minor shortcomings of the nF3 150.
The matter at hand is when exactly are we planning to see these boards? I've seen engineering sample reviews, reference boards, speculation, etc, over the nF3 250 for the past few months, yet I have yet to see a board in the market. The date simply seems to be getting pushed farther and farther back. See the 250 as a minor revision only to the nF3 150, not a whole new chipset, because there only minor differences between the two.
939 is another story.
Another note: It seems like the PCI bus and the AGP bus are synced with one another. Even if I don't set the BIOS to 201 or higher for the automatic lock (which I can't), I have no problems in manually adjusting the AGP bus rate from ClockGen, which also seems to allow me to control the PCI bus rate concurrently.