One more note on your throttle valve. Butterfly valves do not work well in the almost closed position or the almost open position. For the former the valve is overly sensitive (a small change in position yields a large change in conductance). For the latter the valve is under sensitive (it takes a big change in position to appreciably change the conductance).
You can test your set up by manually setting the valve at about 1/3 open i.e. the butterfly at about 30 degrees from closed and seeing how much gas flow it takes to control at your desired pressure setpoint. Then increment to 2/3 open (60 degrees from closed) and see how much flow it takes to get to your setpoint. If you can then determine that your pressure/flow is ok somewhere between these points, you are probably ok. If the valve has to open further or less you are getting into a gray area of control.
This is a very flat footed approach but it'll tell you something about how your system will work. Once you've done this and if it looks fine, then kick in the PID control and adjust the gain (proportional) and lead (derivative) parameters to get the proper control stability. And, do this with air, nitrogen, hydrogen or something else that's cheap. Don't dump your D2 with this set of experiments.
The system I'm setting up will use a 10 sccm (N2) flow controller coupled with a 1 Torr manometer. Under the chamber is a manual butterfly. I can open loop control by partially closing the butterfly and adjusting the gas flow through the MFC's setpoint input until I get the correct pressure. I can close loop control by connecting the output of the manometer to the aux input of the MFC and then delivering a pressure setpoint to the MFC. The only control system latitude is the gain pot on the MFC. This only works with MKS 1159 and 1179 MFCs because of that special aux input.
Steve
Created on Friday, April 13, 2001 12:05 PM EDT by Steve Hansen