I DID IT

grim

Can you expand on your plumbing arrangement a bit?

Water source --> Valve --> Product Condenser --> Reflux Condenser --> Drain

And the probe in the RC.

I've never run this way, but I really wonder if the variable pre-heating of the coolant out of the PC is going to cause the controller to have a tough time. As the RC coolant input increases in temperature the system needs to respond more aggressively. Or, you might hit a point where you want to increase boiler power, and need more PC knockdown, but the only way you can get that is by dropping the RC temp.

For kicks, I would try running the RC on it's own loop, and just used a fixed flowrate through the PC on a second loop. Once you get that dialed in, then you can try the single loop approach and have a solid starting point. I would think it would be possible, if the sizing and flows were balanced appropriately, that a single loop approach might work.

Unsensibel

Plumbing arrangement is exactly as you're describing.

The PC temp is not a factor in the tuning (yet) - I'm boiling water and am tuning for full reflux (~50% power, 35C set point). Once the PID is tuned so I can introduce power (50% <-> 65%) and set temp changes (35C <-> 40C) without issues, I want to see how much the PC temp out is disturbing the loop.

If I get a good PID tune on RC but it's acting up under product take off, I agree that separate plumbing is necessary. However, I'm running with the same plumbing set-up in manual mode & had no issues yet where I would need additional cooling to PC. Usually, with a RC temp out @ 65C, my PC out is @ 45C and distillate temp at 18C.

FloridaCracker

Maybe it's been answered and I missed it, but why is there a T/sight glass above the RC?

Unsensibel

It hasn't ;)

It's a copper packed section - my interpretation of a catalyst of the big rigs.

CothermanDistilling

I am also in the camp that believes separating PC and RC is the way to go... granted, with efficient PC and RC each on their own control, water flow pretty much stays the same for a given input KW and cooling water temp.

I don't wish to be mean, but if you go back 3-4 years in this forum, this is repeated over and over. You are only causing yourself headaches by trying to have both run in the same loop!

I would run the product condenser parallel to the RC, and maybe use a needle valve for the PC(although I will claim the use of a Danfoss Valve is the greatest thing since sliced bread for a PC, it is pure magic at eliminating effort spent on cooling water)

OK, PID tuning... I did a lot of work on the Sestos, but now going to try to redo some of it so you could save a few pennies on another controller. Please, please, please do a write-up on a separate thread with detail on your controller, valve, and system. You can feel free to plagerize anything I write about the MyPin or sestos to write that write-up.

AL1/AL2

you can use these to automatically turn on the cooling water pump and shut the system down if you tinker

Offset

do not use

Input

Pt100

P - Proportional Band (%) Range 0.1-3600

I stole this from Auber controller, it explains it very well, read this 10 times or whatever until you can recite it to others before touching the buttons:

Proportional Constant (P): P is also called the proportional band. Its unit is the percentage of the temperature range. e.g. For a K type thermocouple, the control range is 1500 °C. P=5 means the proportional band is 75 °C (1500x5%). Assuming the set temperature (SV) = 200. When integral, I, and derivative, d, actions are removed - the controller output power will change from 100% to 0% when temperature increases from 125 to 200°C. The smaller the P value is, the stronger action will be for the same temperature difference between SV and PV.

I - Integral Time Range 0.1-3600

From my thread:

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.

From Auber:

Integral time (I): Brings the system up to the set value by adding to the output that is proportional to how far the process value (PV) is from the set value (SV) and how long it has been there. When I decreases, the response speed is faster but the system is less stable. When I increases, the respond speed is slower, but the system is more stable.

D - Derivative Time Range 0.1-3600

from my thread:

Sestos calls this t, or 'Hysteresis time' for some reason.. default is 120, I have it at that.

From Auber:

Derivative time (d): Responds to the rate of PV change, so that the controller can compensate in advance before |SV-PV| gets too big. A larger number increases its action. Setting d-value too small or too large would decrease system stability, cause oscillation or even non-convergence.

Ctl - your manual says set it to zero for 4-20mA, my research says when set to zero it is actually 70, I would follow the manual and set to 0 since you are using 4-20mA

Auber looks to call this 'ot':

Cycle rate (ot): It is the time period (in seconds) that controller uses to calculate its output. e.g. If ot=2, and controller output is set to 10%, the heater will be on 0.2 s e c o n d and off 1.8 second s for every 2 seconds. Smaller ot result in more precision control. For SSR output, ot is normally set at 2. For relay or contactor output, it should be set longer to prevent contacts from wearing out too soon. It normally set to 20~40 seconds.

Unsensibel

@CothermanDistilling I'll certainly write up my summary once done. That's the least I can do for all the knowledge I'm gaining here. Working through the honey-do list right now (which killed my Saturday) but think I'll go back later tonight. I'll certainly will take into consideration you're and @grim 's advice on separating PC and RC. Since I'm on all quick connect, this will be an easy change (but I might just try after getting the PID tuning done to run them in parallel ;) )

razerhawg1

I'm about to move my setup to where I have more head room so I can add a packed section. I thought you normally had your RC above the packed section?

Unsensibel

Hi @razerhawg1 my copper packing is to be used as calayst and not for proof increase. The only relfux there is going to be some passive one. I just wanted to see if I can mimic the German stillmakers setup and increase copper contact...

grim

Doing some more tweaking on the PID last night, trying to do trial/error pid tweaking during a run is painful, we didn't shut down until midnight.

@CothermanDistilling - You are still using a p-band of 30? I'm tweaking in the 5-10 range right now. We started around 20 and moved downwards from that point. 5 is way too aggressive, that's clear. We started with a p-band of 20, and the temp swings were just too big.

CothermanDistilling

I think I am still at 30, might have gone down to 25 or 22, will look.

I like it moving slow, I guess I could try dropping it to 20 or lower, because it never, ever overshoots, which is good for testing, that is, I can jog it from 65C to 68C, and I can drain the parrot and refill in the time it takes to move less than 1 degree C, and take an alcohol reading and dilute and taste, then I can find the point at which it comes off of azeo and the point that taste changes...