I have not written about this in a while, because it usually ends in a pissing match; but why not.
As you can see, the problem the OP had 5 years ago is still persisting today. You can not control fermentation by regulating what I'll call "cold side".
I'll explain.
Let's start out with you pitching your ale yeast into wort that is exactly 67 degrees. You then put a probe into a thermowell, which is in the middle of the beer, and set the controller for 67F. The beer warms up naturally during fermentation, so the fridge/freezer turns on. The fridge/freezer gets extremely cold, and given enough time, it will bring the beer back down to 67. The fridge turns off. Now your cold side problem begins, because you have an insulated fridge/freezer filled with ice cold air. Even though the fridge is not running, the wort temp will continue to drop due to the freezing cold ambient air. The energy from fermentation alone is not enough to offset this drop.
The answer is to control "hot side". If you have a controller on the fridge, you can make this process even more efficient (by keeping the fridge around 55F for example), but I have been successful for a couple years now just using a single stage Johnson controller (set to heat). I use a small dorm fridge that just fits a 6 gallon better bottle. I plug in the fridge to function as normal, and keep it on a low setting. Lets call the setting a 2 on the 1-8 scale. I then run a probe from the temperature controller into a thermowell that extends down into the beer. The controller plugs into a Fermwrap, which is held around the carboy with a bungee chord.
I set the temp controller 1 degree above what I want the beer temp to be, and also set it for a 1 degree differential. So if I am fermenting at 68F, I set it for 69F.
The perpetually cold fridge keeps the beer temperature from getting above fermentation temp, while the Fermwrap keeps the beer from falling below fermentation temp. It sounds odd because of the push/pull nature of the operation, but the result is a dead on beer temperature regardless of how active the fermentation may be.
The other nice thing about this set up is how easy it is to ramp temperatures up as fermentation slows. As you see the activity in the carboy slowing, you can just "up" the temp a couple degrees on the temperature control, and the Fermwrap does the rest. I have found that increasing temperatures during the last 10-15% of fermentation makes a huge difference in hitting those last couple points of attenuation.