Sounds familiar. I had the same exact issue and Bill kept saying it was my regulator. Then send me a bubble trap that did nothing. Removing the check valve leaves the system to dump HIGH KH and LOW LOW ph back into the sump. Design has flaws.
I do have a workaround that I think will work with the check valve in place. I have a Y on the effluent line coming from the reactor, before the controller, or in my case, my Masterflex. One leg of the Y has a valve on it, and using this valve, you can drain water (rapidly) from the ACR after a restart allowing the gas pocket to rapidly expand and reach the check valve, but already at a pressure lower than your regulator pressure. This should avoid the issue I described above.
I have not tested this method yet, but this should allow for safe reactor restarts with a check valve in place.
Dennis