I checked the technical docs on SYJ3120-SMO-M3-F solenoid. It appears to have a flow coefficient Cv = .05 . I could not find any mention of SMO option but I don't think it changes Cv.

When I model the piston actuation speed with this Cv, I get about twice the extension speed for a 1.1 lb pull against gravity. The model integrates the mass flow rates , applies the natural gas law to determine pressures , and calculates the force time histories.

If I adjust the Cv down to match the test data, a factor of .29 reduction is required (Cv about .015).

So, my question is this: Is there anything in the pneumatic system that would restrict the flow rates more than the solenoid spec?

The model assumes some pressure losses for the air tubing and that the cylinder input port doesn't provide significant flow restrictions relative to the tubing diameter.

I also got some additional data from catapult testing to estimate the energy delivered by two pistons. I can match this data fairly well with the .29 reduction in flow rate.

I plan to post the model eventually, but this remains a sticky point for me.

Thanks for posting the component Cv specs. A few were not available from SMC web site but I took a guess at them and approximated the system Cv as follows:
Comments: 1/Cv_sys^2 = sum(1/Cv_i^2) where Cv_i are component Cv's
4-Jan-10
System Cv Calculation (test config)
Part........Number,.. n.. Cv.. n/Cv^2
Pressure reg......... 1 0.111 81.0... no spec
Reservoir Fitting..... 1 0.311 10.3... no spec
tire pump fitting.... 1 0.083 144.... no spec
T fitting.............. 2 0.356 15.8
Solenoid...............1 0.050 400.0
Valve fitting......... 3 0.078 495.9
Cyl fitting............ 1 0.150 44.4
Tubing................ 1 0.110 50.4... assumes .11 Cv per meter
.......................... sum 1241.895
.......................... Cv_sys 0.028

So the Cv_sys of .028 gets closer to the .015 that I need to match the test results. The valve fitting and the Solenoid dominate. I would have thought the tubing would have contributed more. (I assumed a meter length for the calculation).

I suspect there are a few more contributers to system pressure losses that need to be added in somewhere. I couldn't find anything in the cylinder spec re flow coefficients but I'll bet there are some pressure losses across the cylinder input manifold.

I measured the Coulomb friction losses at about 2 oz which should be negligible.

So, I'm still stuck with a .5 fudge factor but that feels a little better than having a .3 factor :)

Any more ideas?