Smart Humidor Build

[morganti]

This might turn into a very long post, but I wanted to share this build with everyone in case you are thinking about doing something similar. I'm going to keep this very high level, if you have questions, just let me know. First and foremost, I am just a marketing nerd that knows how to build webpages and code. My garage turned into a workshop a few years ago and I do as much woodworking when I have some free time. 

Background
I have 3 humidors in various parts of the house and recently my daughter started to use one as a step stool, and I decided that I needed to consolidate everything into one location for many reasons. The second reason why I wanted to take on this project is because I wanted to make my humidor fully connected and smart. I want a system that I could check in without having to be at home, it needed to be fully automated. The other reason for building vs buying is that what I wanted in a humidor comes at a high premium. I looked at the Gerber One, but those things are way out of the budget.

Round 1
Around May I started out modeling the first build around the principals of the Gerber One, and built a cabinet that was going to be a piece of furniture. I had an idea of the basic build, but had way too many competing ideas running through my head, daily design changes, new features to be added. Long story short, things were spiraling before even getting started. So I took a break, during which time I built a 500 sq ft deck, and dislocated my kneecap playing soccer. That was my summer...

Round 2
A few months ago I started to renovate my office and decided that I was going to make a built-in humidor instead of a piece of furniture. I started to pair down the list of features and simplified things significantly and started to map out the automation process to help with the planning process. This was the first concept I put together.

The brains of the system
Once I had the concept in mind, I started to look for wifi enabled micro controllers that could make my vision a reality. After researching many different products and options, I decided to use the Particle Photon. This lightweight dev board  was perfect for what I needed. It was built on Arduino, and designed for IoT projects. I got two Photon's, one to use for the sensors, and one to control the humidification and fans using wifi enabled relays. What was needed now was a whole lotta code and connecting the sensor data to a web service.

I decided to use AWS for all of the data, API connections, and configuration. Once again I needed to map out how all the services would need to connect to one another. This is why I chose to use AWS, it was one platform and I didn't need to patch many services together. Below is very simplified mapping that helped to break down the coding in sections. I tackled one element each week essentially. All said and done I have about 7 individual programs, 5 triggers, created 15+ API endpoints, 6 or so webhooks, error logging, and even SMS alerts.

In order to make all of this possible, I had to create a user profile for myself and set my tolerances, sensor calibration values, and ideal humidity settings. With all of this information I created a program that figures out when the fans need to kick on to recirculate the air. I am further refining it and had to take it offline.

 

The Build
Now that I had a plan in place, I started writing a whole bunch of code during the evenings and testing things out. On the weekends I would work on my office and started to build the cabinet.

I didn't take many pictures at this stage, I just pushed right through. I essentially built a box within a box within a box. The exterior plywood shown here is a no-VOC plywood with Maple veneer. I lined the entire cabinet with 6mil poly (the stuff used as a moisture barrier in home construction), then added blu-wood, a moisture resistant treated plywood. Blu-wood is very safe to use, and used in wet locations like bathrooms or basement construction. The idea is to use the poly as a way to seal the cabinet, the blu-wood allowed me to glue the Spanish Cedar. All in, I lost about 1.5" of interior space.

Lined with Spanish Cedar. The thing in the top corner is actually the sensor in prototype phase.

The electronics of the humidor are all mounted below. I didn't want to take up too much internal space with all of this. Here's what everything does:

1. Main power. This is used to power the wifi unit of the Relay Shield and provide AC power for some of the relay circuits.
2. The sensor unit. All Temperature and Humidity readings are sent by this unit and connect to #5
3. This is a junction box that splits the AC current to the appropriate relays.
4. The relay shield that powers the humdification, fans, and any other electronics. In some cases the relays are powering 12V DC for the fans, and in other cases the relay is powering an AC plug The bottom outlet is powered by the relay.
5. Sensors are connected using Power Over Ethernet PoE. I have about 500' of ethernet cable and a few hundred ethernet ends so I decided to use that for the sensors. This allows me to scale the number of sensors without having to take anything offline, plug and play. I just rig up a new sensor, and plug it in. Instantly readings come through. It's only 3.3V so it's very low power. This was the trickiest part of the programming, how to scale the unit with additional sensors. 

All of the electronics are connected through a GFCI outlet, if even the slightest short circuit occurs, everything is shut off by at least 3 failsafes since there is about 2L of water above the electronics. All of the electrical components are rated for outdoor use. I also need a few outlets for other aspects of the office remodel.

Humidification
I didn't want to buy a cigar oasis for a number of reasons. I know they are great units, but it just didn't fit my criteria so I started to think about building my own unit, and one day I got some inspiration. My daughter had a bad cold and the doctor suggested getting a cool mist machine to help her at night, he described it as a humidifier that saturates the air in a room with cool atomized mist. As soon as I saw it working I had to learn about the system. I quickly found what I needed and built a humidification unit.

It's hard to see, but there is mist being funneled out of the vent and pulled in by the fan. Once the humidity drops below my threshold, the unit cycles up for about 1 minute, and jumps the average RH by 10-12% quickly. At the current time, it takes about 3hrs for the humidity in the unit to hit my lower limit. I need to add some beads in the roof to help keep the humidity stable for longer. There is also a door open sensor that I built using Home Depot parts and a few ft of wire. If the door is open, the humidification will not turn on.

Where I'm at now...

With cigars. There is tonnes of room for expansion, and I still need to add in one more humidor worth of cigars. I just need to build another shelf and re-align these shelves.

The dashboard is pulling in live data can update every minute or so. Right now I have it set to track the last 5 hrs, but I can adjust that to show as far back as I want. The bottom values are the sensor readings, I only have 2 sensors hooked up currently. The chart is actually a Google chart that is plotting the data from AWS, the humidity value is an average of the two units. The status is actually dynamic, so if the door is open, the status will change, if any element of the system that I built is down, it will indicate an outage, if water runs out, same thing. The fan cycle is actually disabled at the moment, so I am just pulling in the last reading value for the time being. Water consumption is another program that calculates depletion based on usage.

Next Steps
I am building an air filtration unit for the roof of the humidor that will pull air in, run it through a 3M filter and then circulate it around. I need a weekend of free time to finish this up. I have 4 more sensors coming in sometime next week, and once those come in, then I will rig them up.

I have to make a few wifi sensors for a friend of mine, so I need to create a nice enclosure for the sensor and the wifi unit to be housed in. Once I get that in place I will apply those designs to my build.

For those looking to build something similar
Make sure you have things well planned out and understand how the data will flow from humidor to your data warehouse. I had the intention of building this to meet my needs, but started to think about how it would be built if it were an enterprise solution. This meant that I had to think through lots of scenarios and possible circumstances that could arise, as a result, the data component is totally scalable now. Right now I have a database of 20,000 readings in there, and the data gets delivered lightning fast. Make sure your data platform is solid.

More to come as things get done.

[Williamos]

Wonderful post. Wonderful project. Thanks for taking the time to share in such detail and seriously, congratulations!


Sent from my iPhone using Tapatalk

[Sean3]

Well done Morganti, looks like you had some fun with this project!  Logic circuits, data logs...Mr. Piggy should enjoy this post.

[Lucifer Morningstar]

Very nice post. The first humidor connected to the Internet I've read about.

[strikethebox7]

Strong work! Great write up for a stellar idea!

[Fairlight]

Very interesting project! It seems you have done an amazing job, especially with software development. 2 small remarks:

1) I see the dht22. If you ever need better sensors go with the sensirion ones. 

2) Your humidity fluctuations are quite large. Beads will not help the problem. Beads are slow and your humidifying cycles are frequent, beads will not have enough time to buffer humidity. Additional air circulation may help. A dehumidifying unit will definitely help. Also try to make the differential value smaller if you have room there. 

 

 

[morganti]

6 hours ago, Fairlight said:

Very interesting project! It seems you have done an amazing job, especially with software development. 2 small remarks:

1) I see the dht22. If you ever need better sensors go with the sensirion ones. 

2) Your humidity fluctuations are quite large. Beads will not help the problem. Beads are slow and your humidifying cycles are frequent, beads will not have enough time to buffer humidity. Additional air circulation may help. A dehumidifying unit will definitely help. Also try to make the differential value smaller if you have room there. 

 

 

Thanks for the feedback. 

The reason I used the DHT22 was because I had one kicking around from a previous project and had the arduino code at hand from a previous sketch. I will look into the sensors your mentioned though, one of the sensors tends to have a long startup time on occasion, as a result I get a spike in the readings, with that being said, I added a delay to the reading cycle and that solves 99% of the issues I was experiencing.

One of the issues I currently have with my setup is the location. It's in a basement with a furnace on the other side of the wall, so it gets extremely dry down there which makes it a challenge even with the seams sealed, gaskets around the doors and openings, all holes plugged, etc. The thought behind the beads was to try and trap some of the excess humidity during the humidification cycle, but what you are saying makes sense. I need to finish the updates to the circulation program, at the moment it's working as expected, but originally it caused some unexpected results that were not ideal - fan ran for 12hrs straight.

[PigFish]

As always I hesitate some when commenting on the projects of others, but here goes!

 

First, very interesting and fun to follow. Thanks so much for sharing…

 

I found this interesting for many reasons, some probably different from my peers somewhat, and that is how your (one’s) perspectives drive the solutions and to what end.

 

I have a always attacked humidor microclimates as a problem of function first, and reporting least of all. With automated function, there is always a responsibility of the humidor administrator to maintain the system. Process automation is not typically ‘self fixing’ but a well designed system can cover up short-term failures. I have always battled with alarm functions… I don’t use them, but they do have a place certainly.

 

From my perspective a project is first evaluated at the operational level, built off of sound humidor theory… Will it work is always my first and most important question?

 

Next is an economic aspect, this is because I now do this on an enterprise basis and everything I build is a (potential) prototype for something that may be for sale. Degree of control is an important part the the operational efficiency as well as a marketability variable… This is because there are already competitors and the bar must be risen to eclipse their current function or there is no need (IMHO) for a different name (mine at least) in the market place.

 

Reporting is very important, but it is largely an error correction function. Automation at its heart is designed to reduce many variables and errors are one of the most important. As an engineer, it is my job to foresee the engineering shortcomings and fix them in the testing phase so that continuous errors need not be dealt with in the end product.

 

For example. If one spends more time working on the aspect of humidor homogenization and circulation, there is no real reason that one needs a report on the humidors lack of balance. Fixing balance as an engineering problem, fixing it at the engineering level eliminates balance as a reporting issue. Reporting is pushed from an end user annoyance to the engineering side and dealt with as an aspect of a requirement, needing to be fixed before the project is labeled a humidor.

 

In other words, one may want an email if the system fails, but does not want an email every hour telling him the system is out of balance and a failure!

 

So for advice, engineer the space and the basic problems first. Set up reporting based on failures and lesser so on anomalies that should be noted and fixed with engineering.

 

It is just a different perspective!

 

 

Thanks again for sharing. Great job and good luck with your project!

 

-Piggy

[Luca]

That is bloody awesome!!! Great work morganti!!! I'm jealous!!

[morganti]

On 1/13/2017 at 3:59 PM, PigFish said:

As always I hesitate some when commenting on the projects of others, but here goes!

 

First, very interesting and fun to follow. Thanks so much for sharing…

 

I found this interesting for many reasons, some probably different from my peers somewhat, and that is how your (one’s) perspectives drive the solutions and to what end.

 

I have a always attacked humidor microclimates as a problem of function first, and reporting least of all. With automated function, there is always a responsibility of the humidor administrator to maintain the system. Process automation is not typically ‘self fixing’ but a well designed system can cover up short-term failures. I have always battled with alarm functions… I don’t use them, but they do have a place certainly.

 

From my perspective a project is first evaluated at the operational level, built off of sound humidor theory… Will it work is always my first and most important question?

 

Next is an economic aspect, this is because I now do this on an enterprise basis and everything I build is a (potential) prototype for something that may be for sale. Degree of control is an important part the the operational efficiency as well as a marketability variable… This is because there are already competitors and the bar must be risen to eclipse their current function or there is no need (IMHO) for a different name (mine at least) in the market place.

 

Reporting is very important, but it is largely an error correction function. Automation at its heart is designed to reduce many variables and errors are one of the most important. As an engineer, it is my job to foresee the engineering shortcomings and fix them in the testing phase so that continuous errors need not be dealt with in the end product.

 

For example. If one spends more time working on the aspect of humidor homogenization and circulation, there is no real reason that one needs a report on the humidors lack of balance. Fixing balance as an engineering problem, fixing it at the engineering level eliminates balance as a reporting issue. Reporting is pushed from an end user annoyance to the engineering side and dealt with as an aspect of a requirement, needing to be fixed before the project is labeled a humidor.

 

In other words, one may want an email if the system fails, but does not want an email every hour telling him the system is out of balance and a failure!

 

So for advice, engineer the space and the basic problems first. Set up reporting based on failures and lesser so on anomalies that should be noted and fixed with engineering.

 

It is just a different perspective!

 

 

Thanks again for sharing. Great job and good luck with your project!

 

-Piggy

Thanks for the feedback and perspective. Sorry for the late response, lots of family stuff the last few weeks, and work....

I totally agree with a lot of your comments and thoughts. By the time I setup the error handling, I scaled it back massively, if 24hrs goes by and no circulation, or my setup is holding steady at -10% for 6hrs, then an Text message and email is sent. After building the humidor, and testing the electronics, the confidence level went up significantly, and the error handling functions became second priority. 

When it comes to the reporting side of things, I deal with a lot of data in my day-to-day, and I can do a lot of really cool things using Machine Learning and predictive analysis models. With that in mind I decided to collect as much data as I can possibly capture. So far I've been able to really understand where the dead spots are in the design, and what is causing the issue (obstructions, density of material, air flow, glass, gaskets). As a result, I had to adjust the angle of a fan and fixed an initial issue. The second problem to overcome is gravity, the dense humid air drops and the fans can only run so much. So keeping the RH difference consistent in a 40" vertical cabinet is my biggest project now. The data has also shed some light on the relationship between temperature and humidity. Something that you have covered extensively

Once I get the air filtration installed, I think that will help with the vertical RH differences.

[morganti]

On 1/22/2017 at 6:21 AM, Luca said:

That is bloody awesome!!! Great work morganti!!! I'm jealous!!

Thanks everyone for the positive responses. Glad to share with everyone. 

[PigFish]

17 minutes ago, morganti said:

 

When it comes to the reporting side of things, I deal with a lot of data in my day-to-day, and I can do a lot of really cool things using Machine Learning and predictive analysis models. With that in mind I decided to collect as much data as I can possibly capture. So far I've been able to really understand where the dead spots are in the design, and what is causing the issue (obstructions, density of material, air flow, glass, gaskets). As a result, I had to adjust the angle of a fan and fixed an initial issue. The second problem to overcome is gravity, the dense humid air drops and the fans can only run so much. So keeping the RH difference consistent in a 40" vertical cabinet is my biggest project now. The data has also shed some light on the relationship between temperature and humidity. Something that you have covered extensively

Once I get the air filtration installed, I think that will help with the vertical RH differences.

This especially is impressive and from my perspective will give you an advantage to making this or future project successful. You must test, collect data, correlate and then understand.

Water is lighter than air by the way. Just note that. If your bottom regions are wet, it is due to excessive evaporation from your humidifier.

Going forward if you are interested in design ideas, or 'hints' learned the hard way, please post up, pm or email me. Posting is best because then many can learn at the same time.

Lastly, general circulation is, by my present thinking, the least effective form of circulation for humidor use. A ducted 'forced inverse convection' process is the best. Look there for some answers.

I am so glad that you did not take my comments negatively. Some people do!!! Good luck, and cheers! -Ray

[morganti]

1 hour ago, PigFish said:

This especially is impressive and from my perspective will give you an advantage to making this or future project successful. You must test, collect data, correlate and then understand.

Water is lighter than air by the way. Just note that. If your bottom regions are wet, it is due to excessive evaporation from your humidifier.

Going forward if you are interested in design ideas, or 'hints' learned the hard way, please post up, pm or email me. Posting is best because then many can learn at the same time.

Lastly, general circulation is, by my present thinking, the least effective form of circulation for humidor use. A ducted 'forced inverse convection' process is the best. Look there for some answers.

I am so glad that you did not take my comments negatively. Some people do!!! Good luck, and cheers! -Ray

The humidification is what stalled me for so long. I bounced around a concept that would require ABS plumbing fixtures to run in the cavity behind the humidor. The idea was to create a vacuum system to pull the humid air up and vent through the ceiling essentially. The downside to this was the fact that I would have to cut a few large holes, and that would weaken the seal of the unit. I didn't want to swiss cheese the humidor first, and I didn't want to loose all that interior space. I went for the less invasive techniques first. Developed a hypothesis, did some research, tested, and implemented.

Now that I have a proof of concept that seems to work in principal, and the data is almost supporting my hypothesis, it's time to revise the idea and refine it further until it hits the mark. Your comments and experience are really helpful. Hell, I may ditch my home-brew design, and start from scratch next month - who knows what I will learn in a few days.

Do you have some resources for the inverse convection concept you mentioned? I will do a search for it tomorrow when I find some time.

As for the comments being interpreted negatively. Not a chance. Your perspective is different than mine, and I appreciate that. Openness to perspective = progress & learning.

Nick

[PigFish]

6 minutes ago, morganti said:

The humidification is what stalled me for so long. I bounced around a concept that would require ABS plumbing fixtures to run in the cavity behind the humidor. The idea was to create a vacuum system to pull the humid air up and vent through the ceiling essentially. The downside to this was the fact that I would have to cut a few large holes, and that would weaken the seal of the unit. I didn't want to swiss cheese the humidor first, and I didn't want to loose all that interior space. I went for the less invasive techniques first. Developed a hypothesis, did some research, tested, and implemented.

Now that I have a proof of concept that seems to work in principal, and the data is almost supporting my hypothesis, it's time to revise the idea and refine it further until it hits the mark. Your comments and experience are really helpful. Hell, I may ditch my home-brew design, and start from scratch next month - who knows what I will learn in a few days.

Do you have some resources for the inverse convection concept you mentioned? I will do a search for it tomorrow when I find some time.

As for the comments being interpreted negatively. Not a chance. Your perspective is different than mine, and I appreciate that. Openness to perspective = progress & learning.

Nick

... yes Nick I do!

This is a mock-up of my Generation 12 cooler, which is now running in the shop.

Too cold to test actual cooling cycles, but a solid surface humidor must necessarily dehydrate or it won't work. It is about 55F in the shop today and I am polishing up some logic settings for a couple of customers that are waiting on the build. The humidor is heated by the way and the approximate 1F delta you see on the log is the result of the heater working and range during the heating process.

Ever since my video review, this box has about 3 boxes of cigars in it. My theory is that one cannot claim a humidor is a success, not a fully controlled one anyway, until they can master control over the 'space' without the need for masses of hygroscopic material to support function. Frankly, I can tweak this further, but I am more interested in a working humidor, not a data log show piece. What this means is that the function must be robust and not plague the owner with unwanted fussing and fixing. While I often will change my temperature up a little in winter to maintain the 70F that I use as a set point, I don't like to have to do too much tinkering with logic once a humidor is set up.

This humidor is the one in the video. The circulation is based on the inverse convection theory.

Cheers! -Ray

[Porch Smoker]

Whole lot of over my head kinda stuff happening here. But awesome work.


Sent from my iPhone using Tapatalk

[morganti]

On 1/24/2017 at 0:06 AM, PigFish said:

... yes Nick I do!

This is a mock-up of my Generation 12 cooler, which is now running in the shop.

Too cold to test actual cooling cycles, but a solid surface humidor must necessarily dehydrate or it won't work. It is about 55F in the shop today and I am polishing up some logic settings for a couple of customers that are waiting on the build. The humidor is heated by the way and the approximate 1F delta you see on the log is the result of the heater working and range during the heating process.

Ever since my video review, this box has about 3 boxes of cigars in it. My theory is that one cannot claim a humidor is a success, not a fully controlled one anyway, until they can master control over the 'space' without the need for masses of hygroscopic material to support function. Frankly, I can tweak this further, but I am more interested in a working humidor, not a data log show piece. What this means is that the function must be robust and not plague the owner with unwanted fussing and fixing. While I often will change my temperature up a little in winter to maintain the 70F that I use as a set point, I don't like to have to do too much tinkering with logic once a humidor is set up.

This humidor is the one in the video. The circulation is based on the inverse convection theory.

Cheers! -Ray

Thanks for the video, makes sense! What was confusing me with the concept was all related to heat. To me, convection is all about heat transfer, so I was wondering if your system was dependent on a heat source. On a side note, I have been thinking about adding a heater to my build to help with stability, but haven't found a good option yet.

Since I still have to fit the fans in the air filter, when I do, I am going to force the air through the front of the humidor to help create the inverse convection cycle. Glad to see you are also using industrial grade fans in your builds, and not using those $3 cheap fans! 

I was toying around with the fan cycles to circulate on a timed basis (using an overly complex calculation, not just on an X minute trigger), and the results were ok in terms of keeping consistent RH from top to bottom. Based on a few points observed in the video, I decided to change this method to turn on the fans when the difference in RH falls outside of a set threshold. Now I have RH consistency within 1% top - bottom. Hope to have the filtration unit complete this weekend, as I am really curious to see what effect it will have on everything.

[PigFish]

I find no reason to halt fan use. That is just my take.

The inverse convention comment was just that, taking the natural action of convection, and forcing it past, or inverse of its natural state.

I do build a timed fan module. This is it... It is for people who don't have some other form of logic control and therefore it stands alone.

If you are going to use a timer, I suggest at least a 25 to 30% duty cycle, and then not a long period cycle, but a short period one, or a high cycle rate. I typically set these up for people to run 150 seconds on and 450 off. That way the humidor is not really stagnant for any long periods of time.

Cheers! -R

[Jimmy_jack]

I'm just here for @PigFish comments. ?