See diagragm.

When I had the test plates, I flowed each plate on different ranges and at different pressures to cross calibrate each range ( 5 ranges total). I have determined approx flow values for each range from testing and comparing to other benchs. Not exact, but has been close enough and very repeatable. When I tested the test orifices, they didn't flow the same on different ranges using my range values. The numbers were 15-20cfm off in some cases. For example with the "300cfm" plate I got 239 on my 354 range at 28", 252 @ 15" on my 248range converting back to 28" and as much as 288 on a 298range at 28" using the extra sliding orifice and the smallest rotating 50cfm orifce.
I have been using this bench for 15+ years with virtually no changes. With a head on the adapter and flowing a port, I have always been able to test on a higher range or lower range and get very close to the same CFM number using my range values, usually less than 2 cfm difference.
When testing the plates I removed the adapter and flowed them right on the top plate. I think the problem may be with how close the higher velocity air thru the orifice is to my internal orifice and the diffuser discs. Possibly causing some backpressure to alter the flow. When a head is attached to the adapter, the air is slowed down thru the cylinder before reaching the top chamber.
I'm not ready to tear it up and start over, I am still very happy with how it works. I am just curious what you guys think the problem might be.
Rick