Jim Miller wrote:

I guess I haven't figured out how the controller makes anything safer. If
the probe goes out of calibration it could cause the controller to call for
more CO2.

The cure for CO2 end-of-tank dumps is a needle valve working against a
pressure head.

To be clear, I'm in the "run CO2 24 hours/day and your pH swing will only be
0.2" camp.

However, about 3 weeks ago my dual gage regulator-needle valve system ran out of
gas. It did an end of tank dump. pH dropped from about 6.4 to a bit under 6
before I noticed fish behavior and turned it off.

So a needle valve does not prevent an empty tank pressure dump.

Interestingly, without any injection, pH only rose 0.2-0.3 per day.Cutting
surface agitation REALLY helps.