Another clutch question

[kdl]

Car is an RF98 honda (converted 2 liter) with a tilton 1000 series HRB and stock Honda clutch and pressure plate. The clutch/slave/engine has run flawlessly for 3 years in this car, and for 4 years in another.
With the engine off, the slave cylinder operates exactly as it should, engages and moves the diaphragm spring tines and is pushed back by same. All settings have been checked, no overtravel, TO bearing to tine clearance as per Tilton specs. This can be done all day long.
With the engine running it is a different story, press once on the pedal, release but there remains a constant force on the tines, the clutch has been disengaged. Turn off the engine and the forces on the diaphragm spring slowly relax and it returns to its static position.
This is what has been done so far
Replaced master cylinder with a rebuilt
Replaced line from cylinder to bell housing (some potential kinks)
Replaced clutch plate and pressure plate with new
Swapped TO bearing/piston with an alternate TO bearing face
Installed new complete HRB unit, mount, TO bearing, piston
Replaced master cylinder with another rebuild
We’ve run 2 weekends like this, getting by through towing the car up to the grid, pushing to get it rolling, putting it into first and racing (thank you hewland for a clutchless shift). Once on the track everything is performing as per normal.
Any thoughts or suggestions are greatly appreciated
Keith

[BeerBudgetRacing]

Seems there is a new friction added because of the centripetal forces.

So, as I read this, if you press the clutch pedal on track it won't re-engage - sticks open.

If you cycle the gas on-off-on several times, does it return? (Thinking the friction might be released.)

How clean/lubed is the input shaft? Does the disc slide on it easily? Is it tweaked at all.
It might be getting cocked a bit and not sliding back to engage the flywheel.

Disc backward? So too much is engaging the input shaft?

Just thinking out loud.....

[kdl]

we shimmed the bellhousing away from the engine by 0.100" and with the engine running everything operates correctly. So what have we counteracted by shimming?
Keith

[kdl]

so over the winter we reshimmed the chassis/engine to eliminate a dog tracking issue. Short version, we returned to the original shimming and the problem disappears. So we have very, very carefully reset the shimming and in the shop everything is working as it should. Racing this weekend so that will be the acid test.
Keith