Be part of our community & join our international next generation forum now!
Wiring a PID controlled mechanical ball valve for constant output temperature on a Reflux Condenser.
By Michael Cotherman
Cotherman Distilling, LLC
24VAC, at least 10VA (volt-amps)
I had an old doorbell brick laying around that worked perfect.
$30 on eBay (a comparable omron is several hundred dollars)
1PCS 0-400 ℃ 5M Type K thermocouple 150mm probe M8 threaded @ eBay
this one has a 5mm probe that fits inside my 1/4" thermistor thermocouple thermowells very nicely.
$8.50 on eBay
(1/2" 2-way non spring return)
$40 on eBay
500ohm 1/2w (I used 510ohm, 1/4w)
old stock from 20+ years ago
'Carlon' 4x4x4 plastic box
~$8 at Home Depot
I marked the outline of the PID opening required using the backing ring. I then drilled a hole in the backside of the front cover with a step drill and used a jigsaw and file to get a nice square hole. Installed the PID into the cover, slip backing ring over, and tighten screws snugly.
I drilled 2 more 1/4" holes in the back bottom, one for the valve cable, the other for the smaller thermocouple cable and the power leads.
Route the valve cable, power cable, and thermocouple cable into the box. FYI - the valve needs to have both DIP switches switched to the UP position, for "Reverse Acting" and "2-10V".
Connect the thermocouple wires to terminals 3 and 4, observing polarity.
Connect the 24VAC power supply wires and the power leads of the Johnson valve to terminals 9 and 10 of the PID.
Connect the grey wire of the valve and the resistor to pin 8.
Tape off the orange wire from the valve.
Connect the other end of the resistor to terminal 6, also run a black jumper wire to the terminal that has the black wire from the valve.
Resistor notes: if you wanted exact value and more power handling capability, 2 1K ohm resistors in parallel would work.
You may have fun fitting all the wires in the terminals, I added a blob of solder to the resistor lead to help the terminal grab it.
I tried Autotune, it went for 10 minutes, and did not seem to work well.
P - proportional - the further away , the harder it tries to correct... I chose a small number, like 30.
I - Integral - Sestos calls this M50 for some reason.... this parameter pushes harder the longer you have been off of the target temp.
Because of the time lag between opening the valve and the outlet temp reducing, compounded by the slowness of the valve, the Integral value has to be pretty big (it is a 1/i, so a big i means small affect.) I ended up using the factory setting of 1000, may be able to go lower.. I tried to turn off the integral altogether, but as predicted, the steady-state always had an error, when set to 75C, it went to 78C and stayed there.
D - Derivative - Sestos calls this t, or 'Hysteresis time' for some reason.. default is 120, I have it at that.
Ctl - Control period default is 4, I have it set to 10
the Ctl value has been noted by some on forums to be screwy:
Ctl Period 40 = 40s 20 = 20s 10 = 10s 8 = 8s 6 = 8s 4 = 8s 3 = 8s 2 = 70s 1 = 70s 0 = 70s
Sestos look amazingly similar to Auber (PDF), they have the similar default values, and this helped determine that t was in fact derivative.