StillDragon® Community Forum

Welcome!

Be part of our community & join our international next generation forum now!

In this Discussion

Dephlegmator Temperature Control

edited December 2014 in Usage

Hi folks! I was wondering if/how you guys control your deph temp. I went by the book for the first time and thought it uses an awful lot of cooling water for the equalization and during the run. Having read some articles from the Austrian distilling institute, it seems to suggest that the optimal temp control seems to be around 55-70 C. Now I'm running a mix of leftover wines and am trying the following: Flush all condensers with fresh tap water @14 C in my area. Run @ 100% until all caps are bubbling. Trickle deph water and let it creep up to 60 C. This will let slowly the heads through. Collect heads to taste and then start collecting additionally in small increments. Control deph temp for the run @ 60 C

Did anyone already play around with different temps? What temp do you like for different types?

«1

Comments

  • Quite a few on here do that, I just installed my first automatic valve, and I recirculate wide open and the water in to water out only raises 2 degrees or so when equalizing with 7700W-11000W on a 8" 4-plate 380L system .

    During the run, I use 11000 W, I set it at 70C and get fairly high proof (180) from 5% still charge.. if set it to 75 or 77, he proof drops and I get more flavor, but I am still learning, only about 10 runs total, 3 automated... when I do a vodka run, I will likely have to lower the temp to stay above 190 proof.

    my water usage is about 1/3 to 1/2 of the 250 gal tote.

    I forgot to post my build, I have now done so.

  • edited December 2014

    I'm not sure how you could ever come up with an "optimal temp" for the depleg. The operating temp of every still can be different due to differences in construction of the depleg, the size, the surface area, the efficiency, the power input to the boiler.

    Not to mention the reflux ratio you are trying to hit.

    You are going to have to figure out your optimal temp yourself.

  • @CothermannDistilling: I like following you posts... so much to learn ;-) To me, the automation is the second step. First I need to find out what to control it to and then work it. Looks like you got a steal on the prop valve. Need to monitor eBay better.

    @grim: Technical differences are totally understood. However, I believe as well that there's a grain of general knowledge that can transfer out of it (ie. if someone runs fruit, grain and sugar, are all the same setpoints used?). Not expecting anything out of this thread other than maybe some pointers on how others are doing it. After all, the experimenting is what it's about

  • Hi @Unsensibel and Welcome!
    On your first few runs I suggest you not use any thermometers of any kind. Learn by using your senses of taste and smell as well as visual observation.
    Once you have the "feel" of your rig then install thermometers and note where the temperatures are - only because they will help you to duplicate a successful run.

    Instrumentation is a crutch that can easily be used to short-circuit your greatest distilling tools; sight, smell, taste and touch.

    Yep, I have still automation now on my 5" rig and I love it. I don't have even one simple thermometer on my Baby Dragon and I get great satisfaction when I run it because it is more "intimate"?
    One is for production and the other is just for the joy if it.

  • edited December 2014

    Agree with Lloyd, this isn't about quality, it's about batch consistency.

    Used incorrectly, an automation system will guarantee you a consistently terrible product.

    Trust your mouth and nose above anything electronic.

    :))

    I still hold my ground that trying to compare operating parameters across different stills, recipes, geographies, will only cause you headache. I remember a long long time ago when I used pot still with a tall mesh packed column, there were major differences in the product just based on the ambient temperature, running in the summer with 85F degree ambient temps produced a product very different from one produced with 45F ambient temps. Why? Because there was a major difference in the passive reflux generated by the tall packed section (no active reflux). Hell, there was one day when we were distilling through a northeaster (winter storm) with high winds. Every time a strong cold wind blew through the shop, the output of the still would drop significantly. Given how seemingly arbitrary things like ambient temperature and a breeze through your shop can make impacts like this, I can't imagine how you can compare between two wildly different still geometries.

  • @grim said:

    Used incorrectly, an automation system will guarantee you a consistently terrible product.

    There is a lot of truth in that statement.

  • There is a method of using empirical data to develop your PID parameters. This is something I have used many times but not in this application. It is a process step method. There are a couple of parameters that if you do not know you can make a pretty good assumption. The first two parameters is the low and high temperature of the system. The third is the % change you will make on your controller which should be easy.

    It is time consuming and so you may not be able to use it but this is a good tool to have. Here are the steps.

    1. Put your controller in manual mode and bring it to say maybe 25% of range.

    2. Your temperature will come up and let sit until it reaches a constant temperature and record the temperature.

    3. With your controller make a step change.

    4. Record the % change you made of your controller. This is your %control

    5. While the temperature rises graph the temperature on the y axis and time on the x axis from the time you initially made the step change

    6. Record the temperature where it again levels to a constant temperature. The difference of temperatures is your Tdelta. Draw a line from the y axis to the temperature it leveled off.

    7. Your graph of your temperature will have a straight line until it starts to rise.

    8. Now draw a line on your graph that is tangent to your temperature line at it's steepest change of your temperature. Draw the line from the x axis to the upper limit line you drew in step 6.

    9. Where the line crosses the x axis is To (Dead Time) and where the line crosses the upper limit line is Tu.

    10. This is where you make the assumption of your lower temperature limit using your controller and the upper limit. Tl and Th.

    11. Now you can find the P, Proportional Gain, of your PID, But first you must find Kc.

    12. Kc = (Tdelta/((%control*(Th-Tl))

    13. From here everything has the data you collected above.

    14. Your P = (1.2/Kc)*(Tu/To)

    15. Your integral time, I = (2.0*To)

    16. Your derivative time, D = (.5*To)

    These are parameters for a First Order Plus Dead Time, FOPDT control. The P is dimensionless. The I & D are of course times. In some applications you will use minutes and some hours.

    It may sound a little difficult but try it a couple of times. Once you get the knack of it it's fun. If you have any questions please let me know.

    You may not have the time and product but if you do it works. It's a great tool to have when your controls aren't tuned properly.

    twobits

  • These PID's don't have a manual mode, at least not that I have been able to find.

    Also, they throw a monkey wrench into it because the P really needs to vary with the output, as when the valve is nearly shut, it takes much longer to make a change in temperature than when the valve is more open....

    I took control theory in college and prefer to tweak by the seat of my pants....

    I have lowered my integral form 1000 to 800, and raised the minimum opening from 77 to 78 (7.8mA) and reduced the differential from 120 to 100...

    I think my biggest improvement will come when I mount the probe tip inside the Reflux Condenser

  • edited December 2014

    If you think the variable flow rate is the biggest issue, why not grab a 3 way mixing valve and plumb it as a constant flow-rate system? This would also eliminate any stagnant locations in either the RC or PC. Depending on the locations of the valves, though, this type of system would likely be slower to respond.

    I'd draw a picture, but easier to just paste one in. Obviously this shows a heating application, but if you use your imagination and flip it, you can see the approach.

    image

    fig_6_3_06.jpg
    539 x 315 - 17K
  • Well, if your controller doesn't have a manual and automatic mode I don't know what to tell you. Sorry to disturb ya bub. I've been working in the chemical process industries for 21 years now. I thought I'd just try to share a little bit of what I've learned over the years. If I tried to "tweak by the seat of my pants" in a plant I believe they would have booted me in my ass and I don't blame them. I guess your control theory class in college didn't go into practical applications. Funny thing about those theory classes, sometimes they just don't translate into the real world.

    Like I said I just wanted to share something because I've been reading a lot of really good ideas and experiences and thought I could contribute something. Forget that. So long Checking out

    twobits

  • @twobits - I'm interested, but I don't understand steps 3 and 4. When you say step change, what do you mean, is it arbitrary? Do you mean go from 25% valve position to 50% valve position, or 8ma output to 9ma output? I guess I just don't understand what you mean by "step".

    I guess one more for step 5, are you simultaneously timing this with a stop watch - x axis is time?

    I'm going to try to hunt down a completed graph.

  • I could have better explained it, sorry about that.

    First, every PID controller I have ever seen has a manual and automatic mode. It's how you stroke valves to check your 4 to 20 mA signal is reaching the valve or any other device.

    In step 1 the controller was put at 25% of your 4 to 20 output, 8mA. The 25% is an arbitrary number I choose, it usually worked for me.

    In step 3 you are raising the output of your controller from the 25% to maybe 30% or 45%. What will help you determine the change you want to make can be a number of factors, the biggest being you want to see a noticable change in temperature and reach a constant temperature in about 30 minutes. The term "step" is used because if you were to graph your controller output in manual mode it woul be a flat line at your 25% output and when you make the change it immediately goes up to the 30 or 45% and the graph looks like a step.

    In step 5 I didn't make it clear that the x axis is time. When the controller output was changed that became time zero. As the temp would go up 1 degree or so I would mark how many minutes from time zero. At the end I had a column of minutes with the column of temps next to it and make the graph.

    A note: In step 6 where I mention the temp levels off. Take this as where it changes very little. Also, nothing is going to be exact no matter what you do (unless it's having exactly one more drink before you retire for the night)

    I sometimes do have problems with numbers. That "going on 21 years now" I mentioned in my previous post is actually 41.

    I hope I cleared some things up. You make a good point in seeing a graph. I'll get a couple.

    twobits

  • I've got Omega and Watlow controllers, they both go manual.

  • Yes, that is a common difference between $30 controllers and $300-600 ones. I have been playing with the cheap ones for nearly 10 years with my espresso machine and electric brewing, the only low cost manual ones I have seen, are the Aubers; but if you need 4-20mA output, you get the Sestos or the Omega... I would be very happy to find a reasonable cost one that had manual, and even more delighted to see one that worked in Fahrenheit...

    That being said, you could get the deltaT by programming the limits to 25% and 50% and have it go from off to on, which would would be going from 25 to 50%.

    This brings up another issue, extreme nonlinearity... The Johnson valve starts flowing at about 7.6mA, and nearly all the control range (say say 68-90C) happens before it gets to 9.8mA, so we may want to got 25% to 35% or something like that

    I disassembled for cleaning last night, and welded the compression fitting into the RC for the PID control, now it will actually measure temp increase without flow happening (temp decrease still does not happen without flow, but that is not a major issue).

    I think this thread will be very good for many out there, I hope my input is considered constructive.

  • edited December 2014

    Yeah, same issue with my Johnson (har har), in fact, I previously set my controllers to use the 4-20mA range versus the 0-20mA range for exactly that reason, it moves the "crack" point lower in the range. The design is really odd, I agree, but my Belimo valves do the exact same thing. I'm sure it has to do with how far the ball needs to rotate to fully close off against higher pressures, or maybe to allow for compatibility on 0-20/4-20 systems.

    My Omegas do allow for me to provide a correction map, so I could set 6mA output to correspond to "0%", however I realized that what I was doing was going to result in a loss of resolution in the "active" part of the valve range. So instead, I just went back to 0-20mA. During a run, the valves aren't anywhere near 0-10% anyway. I played around a few times and it really didn't make much difference once you were running. 25-75% is really where the work happens anyway. I see the same issue at the top end as well, once you hit a certain open %, you don't see much more flow from opening wider.

    Very interested in your feedback on the impact of the probe location change. My gut tells me it's got to be this way, but I've never run it measuring the output temp, so I'd love to know if my gut is right or wrong.

  • so yours runs in the 25-75% open range... mine only runs in the 25-35% range usign 11kw above 8"CD I guess I can run more power, muahahahaha!

    I will still have my BCS monitor the outlet temp, but the PID will control on the internal temp... (and bought a "5~500Hz PWM Duty Ratio to Voltage Signal Converter Module" oneBay to see if I can make my BCS control the temp, would be nice to have the RC start off controlling to 65C, then ramp to 68C during 30min heads compression, then switch to higher temp during run, then shut the valve during tails...)

  • edited December 2014

    Oh, no, that's on the little rig, usually with outrageously high coolant temperatures. The big still is still pending permits being finalized. Boiler is still sitting around since we need to build a masonry boiler room.

    Our new boiler can push 150kw on the output (641kbtu in, 512kbtu out). Should be interesting to see how hard we can push it.

  • I can push 170F coolant, but very low flow when I want 160proof output at a fast rate from 5% wash... trying to balance stripping efficiency with high proof... but I am just learning... maybe I will try 3 elements full on, 16.5KW during a run and see what flow is....

  • I posted a new video in my other thread.

  • For controllers and other hardware and software there is a web site, www.automationdirect.com, I don't know if you've seen it or not. It's pretty neat if you got the process and bucks to get sophisticated. You can put something together that looks industrial grade with a monitor. Don't think I'll get there anytime soon.

    Using a ball valve for control that is the same size as the line it is serving is problematic. Usual practice is reducing the size of the ball valve at least one diameter size. Between the reducers and the ball valve up and down stream have a short piece of pipe. This eliminates turbulence at the ball valve. You can understand the line of rational that if your valve signal variance is small, the valve is too big. If you are operating on the lower end of your limits that is a key indication your control valve is too big. By reducing the valve size you start using a greater percentage of your valve stroke which means you are getting a finer resolution of your control.

    I tried to find a graph of developing the tuning parameters for the PID controller but couldn't. I thought I may have had one on a USB drive with me but I don't. One parameter that I did not make clear was the %control. The % control should be delta%control. It is the difference of the starting % output and the changed % output in your step.

  • I got most of my switches and circuit breakers from automation direct, and they are good for the name brand stuff like eaton supplementary breakers, but not too excited about the import hand-off-auto style switches, just made of cheap plastic instead of high quality plastic, I now search ebay for used A-B switches, they feel so much more professional....

  • Has anyone used EZAutomation already? Only PLC's but made in America

  • looks nice, but would have to see someone else using them in an instance using temp input and pwm and 4-20ma output before I threw several thousand dollars at them...

  • edited January 2015

    Have some experience with PLCs and DAQ I/O setups, but I don't see them being hugely useful. I have boxes full of tear out PLC gear (read: free), but no real desire to use it.

    That said, what I'm looking for is enhanced manual control, not automation. I want information, I want systems that will manage set points, I want data logging and visualization. Take care of managing the mundane so I can manage the important aspects.

    I don't see the usefulness in executing a program for distillation. Automation for mashing might have some benefits, but there are still so many manual processes involved that even calling it "automation" is a stretch. True automation would include grain loading, pH management, dosing additives, mixing speeds, cool down, etc. A massive undertaking. Temperature control is just a minor minor piece of the puzzle.

    Just my 2 cents.

  • @grim To me, PLC is an affordable way to visualize and log data and make a couple tasks more easy. Not looking for a fully automated process but you can introduce a bunch of poka-yoke's to increase safety with it.

  • this does what you ask for $300.. just lack of rtd input and 4-20mA output keep it from being perfect

    BCS-462 Brewery Control System

  • Facepalm! And thanks to Omega, the 4-20mA control is solved pushing another $130. Have you used it already for data logging? Could also be used for fermentation temp control of paired with a couple of SSR's (thinking of driving 110V solenoid with a long duty cycle in the PID or simple hysteresis)

  • details on the $130, please? I use a $30 Sestos for 4-20mA now.. I have a $15 'converter' coming from eBay "5~500Hz PWM Duty Ratio to Voltage Signal Converter" to test using with the BCS for automated changes (such as 20C RC water for reflux, ramp slowly to 40C for fores, 50C for heads, 58C for neutral, 68C for whiskey)

    Want to do a couple fermenters? glad you have another palm! Get a PowerSwitch Tail and run a piece of cat5 cable to them.. bam, done.. oh, the BCS works best with on/off or hysteresis control for refrigerators, all is tunable.

  • edited January 2015

    Based on Omea's website: Signal Conditioners For Programmable Logic Controllers
    They have a 2 channel PWM to 4-20mA converter. Should allow for prop control using the BCS.
    Need to get my hands on a BCS and start logging temps...

  • I think those need a faster PWM, the BCS is 10hz I believe.

    "Output Period

    This is the period of the PWM. The output will be updated as a percentage of this period. For example, if an output controlled to 20% and the Output Period is 10 seconds, it will be on for 2 seconds and off for 8 seconds. The minimum granularity of the pulse is 0.1 seconds."

    Here is the one I got: 5~500Hz PWM Duty Ratio to Voltage Signal Converter Module Pulse Width @ eBay

    Logging:

    Here is a good thread for runnign a logging server: BCS.client - Node.js client library @ ECC Forum

Sign In or Register to comment.