Descent of temperature, wrong?

Hello everyone, I have a question or problem, I’m not sure. Let’s see if anyone can guide me.
I have an old brewpi system with arduino and raspberry pi + running correctly. A fermentor inside a refrigerator, with a heating system by light bulb (see photo

) . The system correctly regulates the fermentation, varying the target temperature of the liquid minimally (Photo1) . Well up here, although I see that when I make a steep drop in temperature, the target temperature and the temperature of the liquid do not meet the same condition. Always descends with 1ºC difference (Photo2). It’s not a big problem, the temperature is going down, but it seems strange to me. I do not know if it is a normal behavior or is a problem in the configuration of parameters.
Then I write down my configuration in case you can see any error that causes this behavior or something else. I have been adjusting intuitively and I do not know if I have finally left a coherent configuration.

Temperature format=Celsius
Temperature setting minimum= -16
Temperature setting maximum=50
PID: Kp =5
PID: Ki=0.25
PID: Kd=1.5
PID: maximum=15
Integrator: maximum temp error °C =0.5
Temperature idle range top = 0.5
Temperature idle range bottom= -0.5
Heating target upper bound= 0.301
Heating target lower bound = -0.199
Cooling target upper bound= 0.199
Cooling target lower bound= -0.301
Maximum time in seconds for heating overshoot estimator =600
Maximum time in seconds for cooling overshoot estimator =2200
Beer fast filter delay time = 79 seconds
Beer slow filter delay time = 159 seconds
Beer slope filter delay time = 8minutes
Fridge fast filter delay time =18 seconds
Fridge slow filter delay time = 159 seconds
Use light as heater = no
Trigger rotary encoder at every … = Full step

I think the maximum difference between best and fridge is set to 10, which is not enough to do the steep ramp down.

Thanks Elco for the quick reply. I can not understand you completely (sorry I need google translate, my English is very poor). I understand that you mean the PID maximum value. and I also understand that the time the refrigerator has to work before the cut I indicated is insufficient. What I do not understand is the value you indicate to me. Now PID maximum = 15 (you have indicated 10). I do not know if what you suggest, is to set the parameter to “10” or, that must be greater than 15.
Thanks again.

Alternately, for your temperature profile you could just create a more gradual transition from the constant temperature to descending temperature. You have enough cooling power with an offset of 10 degrees to cool at that rate, just not enough to catch up after getting behind after the sudden change.

Hello and apologies for the delay, but I wanted to have new information to follow the conversation.
adempewolff, I understand your idea but I am not convinced that time is the problem. I did not want to say this without doing tests.
The point is that the problem has improved a lot. I have increased the value that Elco indicated to me up to 18 and the result for me is satisfactory. I leave you a new graphic capture of the next fermentation at the temperature drop point. It seems enough to me, although I do not know if you think that it should increase the value. What is your opinion?
Thank you.