Basic starting up


#41

When I edit my beer profile, it will not let me save it. In order to set start time I need to store it as a new profile with a new name.
Is everything as it should be here?

An answer is appreciated. :blush:


#42

My questions seem to remain unanswered! Why is this?


#43

Because I had a very busy week!

1-2. I think you should identify the source of overshoot by looking at the control algorithm values during the overshoot. In most cases, it’s integrator windup. This can be prevented by using a longer Ti so it doesn’t increase as quickly and a higher Kp, so the integrator isn’t active at times when Kp should be doing the work.
Anti-windup also triggers when the actuator is at max, so making sure you hit max by having a larger Kp helps.

  1. Date, Beer Temp, Beer Set, Beer Annotation, Fridge Temp, Fridge Set, Fridge Annotation, State, Log 1 Temp, Log 2 Temp, Log 3 Temp

  2. The defaults changed in the latest version because the low Kp and high Ti that were default caused overshoot in many situations.

  3. No.

  4. It receives the values from the script, which has a cached copy of the values on the controller.

About the profile, what are steps to recreate it? If I try with these steps, I don’t have issues:

  • Create new profile, put some values in
  • Apply the new profile
  • Change start time
  • Apply the profile again
  • Refresh page, profile still the same

The profile is only saved when I click apply. If I click ‘save’ in the dialog, it exits the modal but it is only kept in the browser. When I click apply it is sent to the script. That’s a bit confusing, but clicking apply works for me.


#44

Thank you for your answers.
I am presently trying to tune the PID parameters and I would like to share my - up to now -experiences.
I am trying to apply the Ziegler-Nichols approach

  1. I have disabled the outer loop.
  2. On the inner loops for heating and cooling I set TI to 65535 and Td to zero.
  3. I increase Kp in 6 dB steps and run a step response. I am looking for overshoot and oscillations
  4. I have not found any overshoot yet.
  5. However, the resulting temperature comes dead on the set temperature when I increase the gain. As it should.
  6. My present conclusion is that there is very little excessive stored heat energy in the process, thus we see no overshoot.
  7. Presently I believe it will work best as a pure P regulator, the integral part will be a disturbing factor - possibly causing overshoot in a step response.

Can I inhibit the integral part by other means than setting Ti = 65535?

I am looking forward to your comments.


#45

If you want to disable the integrator you can set Ti to zero.

The integral is to correct steady state errors due to constant losses. Without it, you’ll probably end up at a constant value close to the setpoint if the temperature difference with the environment is large.


#46

Thanx again!

I have the understanding that Ki = 1/Ti. Would I dissable integration by setting Ti=0?

Also, are you able to provide a schematic diagram showing where the low-pass filters (denoted as time delay) are in the loop?

Thanking you in advance.


#47

Yes. That’s correct. There’s a check for zero in the code.
The filters are at the input. Full algorithm here:


#48

Yepp, Ti = 0 seems to work well. :blush:

How do I delete beer profiles? Where are they located?


#49

You can delete them in the UI or in /var/www/html/data/profiles


#50

When I have a stable set temp 20°C, temp outside fridge 16°C, fridge temp increases and cooling engages. I think my culprit is the fan that is on constantly and draws 38 VA.
I see the posibility of steering the chamber fan
a} 230V thru an SSR.
b} Antec tricool powered by 12V and controlled by a 5V from an actuator port.
1} How is this function programmed to work?
2} Do we have PWM control of the fan?
3} Is the fan on only during heat and cool?
4} Can I set any parameters here?
5} Any advice for me here?
Cheers