I did some verification of the original restriction tests, and also relocated the pressure sensor to the other side, the long sweep side. The losses were half of the original values. All this means, what I suspected, is that there is a massive pressure gradient across the mouthpiece x-section. At max airflow (54.5 lb/min, 22 psi boost) there is a range of pressure readings, depending on where the sensor is located. On the inside of the corner, it maxes out at 180iwc, (approximated, the gauge only goes to 150). On the outside sweep, it is 60 iwc. I can only try to average the loss, at about 120-130 iwc, which is useful for the spreadsheet predictions, posted earlier. But for testing improvements, I will stick with the original location.
Now for the improvement. You should really like this. I spent 15 minutes machining the corners as we had done in earlier trials. I did this for a couple of people as well. The loss went from 150 iwc to 120! The MAF went from 54.5 to 57.5 lb/min, a repeatable observation! This should be good for a few ponies and IMO is well worth doing. Oh, I almost forgot, COT dropped 30 degrees also. I can't rule out placebo, but it felt like it spooled up a tad quicker.
So I will definately be going the route of beefing this up to a new proto, and making for some real big changes.