Sorry, things got busy this week. Sorry if this gets long, but to provide context for anyone in the future that looks back at this thread, I'll go into more detail than maybe is needed right now.
Yes, the enclosures get expensive. I think over the last 4 months I have spent $40k on enclosures for some of our production branches. I have had some luck using the plastic enclosures that you can get at Lowe's. I set up an Inkbird (different model) for a fermenting chamber and used a plastic enclosure.
PID (Proportional, Integral, Derivative) controllers work off the error of the control variable which in this case is temperature. To calculate error take the set point and subtract the actual read value. So T-error = T-setpoint - T-actual. Then Output = P*T-error +sum(I*T-error). Note that the integral output is the summation of error over time. The difficulty that we run into then is output due to the integral term spools up and causes overshoot until the error is negative for long enough that the integral error is reduced to zero.
Since we are melting lead, the system should have a long time constant, meaning that it does not quickly respond to changes in the input. This will result in a slow oscillation around the set value. Here is how I would set the system up if it were mine. Keep in mind that the documentation is pretty poor and I may be completely wrong in how I am interpreting the user manual.
I would be concerned about enabling the control as soon as the pot is started. This would give the integral term a huge amount of time to accumulate error and the temperature could way overshoot. But reading section6.24 of the user manual it says, "If the temperature control requirement is undemanding ..... low temperature point = SV setting value - dF and The high temperature point = SV setting value + dF." I think that this means that the controller is only active during the High and low temperature points. It appears that under screen ALP --> dF you can set the dF value. I would set Df to something around 10 degrees.
Then as far as the PID screens.
Control Period: 1 second: Since you have an SSR it can handle higher switching frequencies than a normal coil relay. Normal coil relays should not switch more than once in 20 seconds.
Proportional Band: 200: I would start at a low value and make sure that the I and D terms are set to 0. This should allow you see how the value of the P term effects the pot. Start with 200 and see how well the pot maintains close to the desired set point. If the temperature continues to fall over time, then increase the value in increments of 50. If the temperature starts to run away, then decrease the value.
Integral Time: It is unclear to me why the units on this are seconds. But anyway I would treat it just like we did with the P term. Once you have the P term set where the pot is maintaining temperature (we don't really care that the error is zero, but we do care that the temperature is stable), then increase the integral term to 100. This value should be low enough that the integral error will not cause huge overshoot. Once you have seen how this value performs, then turn the pot off and let it cool by 20 degrees. Once the pot has cooled, then turn it on and see how badly the temperature overshoots or how long it takes to achieve zero error. If the pot overshoots temperature by a lot, then turn the integral term down. If it takes FOREVER to achieve the set value, then increase the term.
There are more scientific ways to tune a PID controller, but I have found that they are often better suited to more complex problems where the risk is higher. For something simple like this, playing around with the values is typically the best approach. The only thing to really be aware of is don't change to many things at once. Change one variable at a time.
Hopefully that helps! Let us know how your progress is. You have me wanting to do this now too just for fun!
