Raspberry Pi Still Contoller

[TimothyChurch]

I've decided to take things to the next level! After getting a few things squared away professionally, I have a timeline for setting up my DSP here in a couple of years. Until then, it is time to start making sure my practices are more on the professional level and start getting all my systems in place.

One of the things I've always wanted to do it to automate my depleg control. I run a modular bubble plate still with up to 4 plates at the moment. As legally in the US whiskey has to be 160 proof or lower, I've been running 2 plates on my flavored product and dialing in my take off to 80%. I normally go through the song and dance at the end of the run of adding some heat and increasing reflux to push down the late tails a bit. While for repeat runs, this isn't bad they have their 6%ish ABV plus usually, a gallon of feints at 65%, the first one of the batches have so little alcohol they need a bit more pampering!

So in comes the automation!

**DISCLAIMER** Just because I know it will be said because it always is. Just because I am automating something, does not mean I'll be leaving the still unattended. I'm simply taking one variable off my hands so I can fine-tune my product!

Here is the rather rough dashboard at the moment

I have 5 temperature probes: the jacket water, the pot, the vapor at the bottom of the column, the dephleg output, and the head vapor temp.
The control points will be the heating element as well as two valves directing coolant towards or away from the dephleg.

Here are my proposed steps for running it with my system.

Turn on the system and hit preheat. Preheat is going to be set just under the boiling point of some of the lower boiling heads for safety. The jacket and pot are run using a cascade PID system so there is no pressure build-up. (System is rated for 15 PSI and I never go above 5 even when cooking corn)

Turn on the water. The coolant loop will feed from the city water. First will be a large needle valve for overall flow control with a bypass valve just in case. The water will flow into the PC and then from the PC to two valves. The valve going to the depleg will be a default open valve with the bypass being a default closed. While normally people run two loops (that's how I have it now), that is typically to avoid constantly adjusting things. The point of this is to be able to constantly adjust the system and the slightly warmed water from the PC allows me to have a slightly higher flow through the dephleg.

Once the still is preheated, it will be switch to manual mode to begin the run and stack the plates. The needle valve will be adjusted till the dephleg output is just over full reflux. At this point, the dephleg target temp will be raised to draw off heads.

Once heads have been collected, it will be dialed in to the desired ABV based on dephleg temp and allowed to run.

At the end of the run, there will be a tails button to switch off the dephleg and finish the run. Once the charge temp shows it to be below a certain ABV, the system will shut down.

While this is happening, there will be charts showing all the temps as well as the ABVs. I'll set these up to get logged for each batch.


Now this is still in development! I'm working on the system using Node-Red for easier programming (my code is a bit rusty). I'd love to get any thoughts y'all might have or features you think would be nice to see. I'd be happy to share the code once I've flushed out all the bugs!

[jward]

I have been thinking about some automation too geared more towards safety then product. Some ideas are a dead man switch and monitoring the temp probes for unsafe conditions. Reading the thread on mistakes made it seems that overflowing the collection jar was quite common. Monitoring container fill or weight could be handy.

[Mr Sippy]

Ambitious

My memory on this is very dim but I believe a similar project was undertaken by a fellow who went by the name of 'amanofconstantsorrow. Pretty sure it was on this forum. His difficulty came when, as the still heated up, the performance characteristic of the stepper valve would behave unpredictably. Not sure if he was able to resolve it. It sounds like you have some experience with PID tuning so maybe it won't be a problem for you, but just wanted to mention it. Assuming, of course, that is part of your design. Good luck with your project.

[TimothyChurch]

I have been thinking about some automation too geared more towards safety then product. Some ideas are a dead man switch and monitoring the temp probes for unsafe conditions. Reading the thread on mistakes made it seems that overflowing the collection jar was quite common. Monitoring container fill or weight could be handy.

While my controller box won't have controls on it, I was thinking about just having one big kill button to shut everything down if need be.

I really like the idea of being able to monitor collection amounts. I thought about seeing if there was a way of calculating it by knowing the ABV of the wash and the product but I think that would just be a long complicated rabbit hole. I know most craft distillers measure based on weight. It could be an awesome option to keep tabs on collection rate, amounts for cuts, and set alarms for separate jars if collecting individually! I'm going to think about this one!

[TimothyChurch]

Ambitious

My memory on this is very dim but I believe a similar project was undertaken by a fellow who went by the name of 'amanofconstantsorrow. Pretty sure it was on this forum. His difficulty came when, as the still heated up, the performance characteristic of the stepper valve would behave unpredictably. Not sure if he was able to resolve it. It sounds like you have some experience with PID tuning so maybe it won't be a problem for you, but just wanted to mention it. Assuming, of course, that is part of your design. Good luck with your project.

Thanks for letting me know! I'll have to see if I can find his project so I can see what roadblocks he hit. I wasn't going to even deal with stepper valves and proportional control for that reason. It is a lot easier for me to think of valves as on-off and just manual control the overall flow.

Currently, I have dephleg control with a simple on-off setup. Higher than the target, all flow goes to dephleg. Below target temp but higher than low threshold both valves are open to slow flow to depleg. If it gets below the lower threshold, all flow goes to bypass. It is one thing to talk about in theory but I'm interested to see how it performs once it is all set up!

[Yummyrum]

Some ideas are a dead man switch

I think this should be mandatory on all stills , not only ones with some attempt at automation .

[bluefish_dist]

Currently, I have dephleg control with a simple on-off setup. Higher than the target, all flow goes to dephleg. Below target temp but higher than low threshold both valves are open to slow flow to depleg. If it gets below the lower threshold, all flow goes to bypass. It is one thing to talk about in theory but I'm interested to see how it performs once it is all set up!

I think you will find that won’t work well. I think you need constant flow. A three step might work, low, med and high deflag flow. I would suggest using pwm with a on/off valve and set a base %, then adjust it up if it’s too hot, down if it’s too low. Allow a bit of time between making adjustments to help keep it in control. Good luck.
I had considered adding controls to my vm to hold a constant temp. That would have been easier. Simply open the valve, then close it in known increments when it hits a set temp. If it dropped below the set point by say 5 deg, then open the valve. Don’t allow More than one adjustment in x min to prevent it from bouncing around too much.

[shadylane]

I'm thinking one way to control a dephleg with automation
Might be to preheat the cooling water...instead of trying regulate the flow.

[TimothyChurch]

I think you will find that won’t work well. I think you need constant flow. A three step might work, low, med and high deflag flow. I would suggest using pwm with a on/off valve and set a base %, then adjust it up if it’s too hot, down if it’s too low. Allow a bit of time between making adjustments to help keep it in control. Good luck.
I had considered adding controls to my vm to hold a constant temp. That would have been easier. Simply open the valve, then close it in known increments when it hits a set temp. If it dropped below the set point by say 5 deg, then open the valve. Don’t allow More than one adjustment in x min to prevent it from bouncing around too much.

I've thought about PWM control but I was worried about the extra stress on the valve. Didn't know if cycling it too much would run the risk of my valve failing in the future.

My hope was that by setting the overall flow rate of the system just above the total reflux, the amount of cycling would be decreased. I've found when running my still that it is only a little difference in that needle valve from total reflux to production! I could set up multiple feed valves to the dephleg each with its own needle valve set at different points but that seems overly complicated to run the system.

I know a lot of larger still are run using thermostatic valves as they over great proportional control based on temperature. The only downside of those would not be having the ability to change the target takeoff abv easily.

[TimothyChurch]

I'm thinking one way to control a dephleg with automation
Might be to preheat the cooling water...instead of trying regulate the flow.

What would you do if you overshot the temp and ran too hot of coolant? Seems like it could be difficult to regulate the input temp of the water but it is an interesting idea.

[Doctaf]

Couldn't you just run a small rc servo coupled to inexpensive valves to adjust your input and output flows based on an set temp value using a pid controller with the "heat" side of the unit controlling the downstream valve?

[TimothyChurch]

I have been thinking about some automation too geared more towards safety then product. Some ideas are a dead man switch and monitoring the temp probes for unsafe conditions. Reading the thread on mistakes made it seems that overflowing the collection jar was quite common. Monitoring container fill or weight could be handy.

It turns out it actually wouldn't be to hard to set up a weight measurement sensor. I can take apart a digital kitchen scale and wire it up. The only downside is that I'd have wiring going to my collection area so I'd need to make sure everything is thoroughly sealed and watertight. OR! I could set it up using a Bluetooth Arduino board and make it send things wirelessly!! That might be a large rabbit hole though.

Really opens up a whole area of features. Being able to measure collection volumes in real-time and even use that to adjust heat input to raise or lower flow rate. Then have the valves keep the dephleg at the correct ABV even with the change in input heat! At this point, I'll need to perfect a baseline system and then do add on features as one-offs to it or else I'll have too many things part way done.

[Doctaf]

The problem with using Bluetooth/wireless is the latency involved with data collection. Your better bet would be to pot your wires with epoxy using an arduino specific load cell

[shadylane]

I'm thinking one way to control a dephleg with automation
Might be to preheat the cooling water...instead of trying regulate the flow.

What would you do if you overshot the temp and ran too hot of coolant? Seems like it could be difficult to regulate the input temp of the water but it is an interesting idea.

There's many possible inputs to the little computer
But there's only a few major outputs
Mostly boiler power and reflux control

Have the water flow constant and vary the power to a dephleg water preheater.
For automation, I think it might be easier to control the water temp going into a dephleg than it would be to accurately control the flow using servo controlled valves.

[jward]

Reading the thread on mistakes made it seems that overflowing the collection jar was quite common. Monitoring container fill or weight could be handy.

It turns out it actually wouldn't be to hard to set up a weight measurement sensor. I can take apart a digital kitchen scale and wire it up. The only downside is that I'd have wiring going to my collection area so I'd need to make sure everything is thoroughly sealed and watertight. OR! I could set it up using a Bluetooth Arduino board and make it send things wirelessly!! That might be a large rabbit hole though.

Really opens up a whole area of features. Being able to measure collection volumes in real-time and even use that to adjust heat input to raise or lower flow rate. Then have the valves keep the dephleg at the correct ABV even with the change in input heat! At this point, I'll need to perfect a baseline system and then do add on features as one-offs to it or else I'll have too many things part way done.

I saw the maker kits for weighing things and thought as you do using a digital kitchen scale would be a good way to go.
Someone here shared this fill detector:https://www.ebay.co.uk/itm/Audible-vibr ... 3421719069 One could probably use circuit continuity for fill detection as long as there is no spark when the circuit connects.

[shadylane]

I have been thinking about some automation too geared more towards safety then product. Some ideas are a dead man switch and monitoring the temp probes for unsafe conditions. Reading the thread on mistakes made it seems that overflowing the collection jar was quite common. Monitoring container fill or weight could be handy.

Having a big catch pan under the collection jars is a good idea.
If anything over flows, it stays in the pan.

[TimothyChurch]

I saw the maker kits for weighing things and thought as you do using a digital kitchen scale would be a good way to go.
Someone here shared this fill detector:https://www.ebay.co.uk/itm/Audible-vibr ... 3421719069 One could probably use circuit continuity for fill detection as long as there is no spark when the circuit connects.

I think I'd lean towards the weight approach. While I know it would most likely be okay, the thought of some sensors being in the actual distillate doesn't sit right with me. I actually was looking through the draw and found my old brewing scale I used for hops! Gives me a free one to take apart and tinker with down the road.

[TimothyChurch]

By the way, I don't know if anyone would be interested but if you'd like to play around with the software here is the GitHub for it.

https://github.com/TimothyChurch/Distilling-Software

As I've mentioned, I'm still in the middle of the development of the base program. Since I enjoy this type of thing, thought others might as well. Feel free to branch off of it and play around with customizing it for your system. It is written using Node-Red.

[Doctaf]

By the way, I don't know if anyone would be interested but if you'd like to play around with the software here is the GitHub for it.

https://github.com/TimothyChurch/Distilling-Software

As I've mentioned, I'm still in the middle of the development of the base program. Since I enjoy this type of thing, thought others might as well. Feel free to branch off of it and play around with customizing it for your system. It is written using Node-Red.

I cant speak for others but im very interested in following this,

[tiramisu]

just a post so I get updates.

[TimothyChurch]

Well, parts are ordered!! I decided to go with RTDs for temperature sensors instead of the typical DS18B20 that are used with Raspberry Pi. It means I will have a handful of converters that I have to wire up but I think the accuracy will be worth it.

I also got a lot further on the software but forgot to upload it to Github before I left for vacation. I've been twiddling my thumbs without having it to work on in my downtime. The latest iteration has a logging section and I'm beginning to flush out an initial entry form with stats about the run (volumes, abvs, feints, etc.) to help the software log things better.

I'm thinking of eventually making that form even more dynamic where you can specify what type of rig you are running and what sensor you have active. That way people who have a different set up can easily use the software. For instance, I run a jacketed rig with a cascade PID setup. For those with direct fire, this doesn't have a benefit so I'd like for the user to be able to specify. That way it would be more customizable per person.

I'd love to hear about what type of rig you'd like to run with it so I can get an idea of other branches to add to the software. It would be great to be able to do various VM, CM, & LM setups on the same software. In fact, now that I say that I think I'll add the coding for the DS18B20s back and just set up a switch to specify what type of sensor you are running!

Let me know your ideas! I'm enjoying the planning and building of this and would like to truly create something for anyone who would like this type of interface!

[SomethingObscure]

I'm very interest in following your progress. I'd be quite interest to see this applied to a ccvn. I would have thought a linear actuator of some kind to move the reflux condenser would be easier to program than controlling flow or input temp on a dephleg.

Do you think you will ( with accurate enough weight measurements) be able calculate output flow rate?

If you wanted to control the coolant temp of the dephleg it would be much easier if you had access to both hot and cold water and utilise a thermostatic valve.

Anyway I'm such a newb here, but just my thoughts.

Keep it up
SomethingObscure



“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.”
-Albert Einstein

[HDNB]

weight varies. i'd go with a float level sensor and motorize from there.

if you find suitable motorized needle valves, share that info and links, i never did find a workable solution.


actually, links on all the parts is always a helpful addition to posts

[SomethingObscure]

weight varies. i'd go with a float level sensor and motorize from there.

if you find suitable motorized needle valves, share that info and links, i never did find a workable solution.


actually, links on all the parts is always a helpful addition to posts

Could you not use a standard valve with a stepper motor directly connected.

Actuated needle values will be pricy

https://uk.rs-online.com/mobile/amp/p/p ... lsrc=aw.ds

“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.”
-Albert Einstein

[tiramisu]

Not an arduino but I suspect you may appreciate the weight component

[TimothyChurch]

I'm very interest in following your progress. I'd be quite interest to see this applied to a ccvn. I would have thought a linear actuator of some kind to move the reflux condenser would be easier to program than controlling flow or input temp on a dephleg.

Do you think you will ( with accurate enough weight measurements) be able calculate output flow rate?

If you wanted to control the coolant temp of the dephleg it would be much easier if you had access to both hot and cold water and utilise a thermostatic valve.

An automated CCVM..... Now that could be awesome! I'm just picturing a large actuator mounted to the top of the column. The only issue I see with that would just be making sure the whole setup is stable. With a 1 meter column for neutral, that would be a lot of weight up high. I'd almost think you'd have to have a way of bracing that to the wall.

I don't think I'll have too many issues with getting the flow rate through weight. Ethanol weighs .79 g/mL and even a cheap live load sensor is accurate to within a gram once calibrated. The measurement would be within just over an mL which is more accurate than my eye can do on a mason jar! Since I have the ABV from the product vapor temp, I can make a simple equation to get the weight per mL of that ABV. The only thing I'm trying to work through is resetting the weight between each jar. I'm thinking have the sensor measure every 10 seconds and if it reads no weight it shifts down to the next jar in the chart. Fun coding to come!

I've thought about controlling the water temp but from an efficiency standpoint, adding more heat into the system is going to just make additional waste. Maybe not a lot on our scale but I want to make sure anything I do is scalable easily!

[TimothyChurch]

weight varies. i'd go with a float level sensor and motorize from there.

if you find suitable motorized needle valves, share that info and links, i never did find a workable solution.


actually, links on all the parts is always a helpful addition to posts

Yea I am going to have to have the program constantly adjust for weight changes based on ABV. Since I'm collecting in jars, I wouldn't want to have to move a float over every time I change things out.

The needle valve is the one thing I'm keeping manual at this time to control the overall flow rate. The ball valves will be just simple on-off setup for now.

I'll definitely put together a part list for the forum here in the next day or two. I was going to wait to make sure it worked first but I guess I can share it either way.

[TimothyChurch]

Not an arduino but I suspect you may appreciate the weight component

Now that is AWESOME! I've seen one that was a basic conveyor belt approach but this is amazing! If it will listen to me ramble on and on about things I might have found my new favorite bartender!

[Mr Sippy]

Hi Tim-- keeping an eye on this one.

If you don't already have it, and can use C code; I have written a routine that equates mol%, abw, and abv based on atmospheric pressure and boiler temperature.
https://homedistiller.org/forum/viewtop ... 97&t=79784
Have a look at the output of the zipfile and see if its useful. The equations (from the literature) were generated under controlled conditions so they ARE the baseline. Then you parameterize to fit your conditions; coefficients, systematic error, etc. My program includes parameters for atmospheric pressure or altitude. Currently working on code for heat loss, mass flow rate, and other instantaneous conditions.

I have a scale wired into a microcontroller. The signal is just a bit noisy. Using certified weights in various combinations of 50-100-500 grams. I recall +/- 1-2 grams. Might smooth out with a moving average... don't know yet. Depending on jar size, control by weight might be iffy. My program output will confirm that. Hope is for a PID power control through the scale. This is a casual project for us so we work on it when we can. I suspect your timeline and energy level register a bit than ours higher but I'll share where I am now.

PM me if you're interested in the source code.

Regards

[RC Al]

Just as an add on, the product weighing idea has been on my mind for a while, consider adding rf id chips to the cut jars so that the tare weight of each one can be accurately accounted for and you can use different sizes with no issues for the software