Peculiarities of Webb-Kelly tank

Serves me right. After being so jubilant about my experimental 29 gallon
Webb-Kelly tank, a problem has developed. The first four weeks I ran the
water return from the Fluval Internal Cannister through the submerged
spray bar and directly into the tank. Then, I switched a small portion of
the flow down the UG tube for a slow RUG effect, to simulate ground water
upflow. Now the problem.

                Water Parameters
               Before       After
KH (TETRA)      2.0/2.5      3.0
pH (TETRA)      6.4          7.2
CO2 injection   x/3 sec      x/2 sec  where x is 1 bubble

The substrate has a layer of fine clay and vermiculite through which the
water passes. The DIY CO2 generator seems to be producing enough CO2 and
there appears to be no leakage in the bottle, tubing, or connection to
the Fluval which is submerged. And the gas seems to be dissolving. But
the pH will not drop and the plants are not growing as well as before.
I'm actually using almost twice the yeast as previously, but to no avail.
The water is not supersaturated with oxygen as before, although the c.
nevillii, c. wendtii, and rotalla are growing well. Thinking that there
might be venting of the CO2 via the fine stream flowing down the lift
tube, I returned to the original configuration and submerged the entire
spray bar. No improvement. Also fiddled with room temperature, new CO2
generators, added micronutrients, macronutrients etc. Surface movement 
remains stable and is negligible. It's in the water!

Now, if you've followed me this far.....

If the clay has a limestone component, the return of the CO2 laden water
through the substrate may dissolve some CaCO3 to form Ca(HCO3)2. But why
doesn't the KH reading show a marked increase as one would expect given
the much higher pH and the increased CO2 injection? I believe the clay
to be a marine clay mixed with some silt. I tested a dry bit with
several drops of HCl. There was no fizzing as far as I could tell.

I understand that phosphates and silicates contribute to alkalinity.
These could be found either in the clay or in the pond tablet fertilizer.
Does anyone know if this might be the answer? Wouldn't the TETRA KH test
also falsely measure these as carbonate hardness and so produce a high
phony KH? Any ideas? The most credible answer receives the Whittaker
prize in stoichiometry.