If you have not already done so, pick up a small dosing pump for use with the reactor. Once you have a pH controller governing the CO2 and a dosing pump controlling the flow through...you have complete, reliable control over the reactor.
Bubble count/rate no longer matters and you can get rid of any needle valve that is controlling flow rate. Both are weaknesses in consistency.
Then...start testing various combinations of pH and flow rate by testing the alk of the output. (Use 1/2 the regular tested amount and multiply the answer by 2 or you will use up all your test kit reagent really fast) IME, all reactors have a sweet spot maximum. If you flow too much through them the net delivered Alk goes down...if you decrease the pH too much the media turns to mush.
Take (Output Alk- Tank Alk) x flow rate = "output number".
ie: (20dkh - 8dkh) x 40ml/min = 480... (18dkh - 8dkh) x 50ml/min = 500... (16dkh-8dkh) x 60ml/min = 480. In these three examples the net contribution to the tank is about the same assuming the tank is running at 8dkh.
Once you've done all that, you'll know for sure if your reactor is simply too small...or just not reaching it's potential.