20 litre paint buckets with 1-2kg of silica gel cat litter just tossed in the bottom.
Stacker is an American maker of expensive, high-end industrial 3D printers. It turns out they also make dry boxes, which they named āF-Boxā, and their FAQ for those makes for an interesting read. Among other things, it sounds as though even without heating you can dry your filaments in a dry box provided the RH inside it kept low enough via desiccation:
Q: What is the ideal RH weāre trying to achieve?
A: IIn general, you want your filament to be as dry as it will go. In the F-BOX itās easy to get the RH well below 10% at room temperature.Letās assume your room temp is 22C and the RH inside your F-BOX is 5%. In this example the dew point is -19C. This is considered quite dry and will quickly remove the moisture from your filament.
F-BOX (Filament Dry Box) - STACKER 3D
Seems like a credible information source to me, yet that info seems to fly in the face of what everyone assumes, which is that you need proper heating to dry your filament spools. 
Who is right? How much slower or faster would that approach be? Luckily, that claim is easily tested by weighing a filament spool before and after putting it in such a dry box. They use molecular sieves in theirs. Maybe it matters? If the Stacker FAQ assertions are correct, then maybe do your bulk drying with a heated source and ambient make-up air until it dries no more, then do the final polish-off drying in such a drybox adequately stuffed with the right kind of desiccant to fully dry it, regardless of how high the ambient humidity outside the dry box might be. If it works, then that would be maybe the most ez-peasy solution of all, even for people in tropical locations, such as where @IslandBill is.
I ordered a couple of their F-Box dry boxes, and as Project Farm would say, āWeāre gonna test that.ā 
1 Like
Heating excites the molecules to move away from where they are. The increase in temperature causes an increase in movement which, for example, forces the air trapped in the hollow spaces between filament coils out and about to be available for dehumidification. The idea is to force the air to move to shorten the time needed to remove the moisture from a spool.
If I want to wait a log time, Brownian motion will eventually force each molecule of water to meet up with a piece of desiccant. The heat is there to reduce the time frame, nothing else. If one has a perfectly sealed container waiting for equilibrium is possible. Nothing some cheap Chinese manufacturer builds is going to form a perfect seal unless youāre an anal engineer like me and does a DIY project.
Itās why Iām only using tiny heaters (22W) to get a certain amount of motion going without some tremendous amount of heat affecting the chemistry of any filament. Itās also why Iām not using time to determine when to stop the process. Iām using an interrupt driven model that notices when humidity reaches a low, an undetermined low, and then stops.
1 Like
@IslandBill Sounds like youāre nearly done then. I subsequently ran across these:
which you could put in-line with your aquarium pump. Interesting how thereās such a sharp line between active and exhausted desiccant:
Presumably you can watch that line move over time and use that information to predict how much time is left before needing to regenerate.
2 Likes
I checked their site for statistics. It wants 10X the flow rate I intend to start with. The heating chamber in my design is cramped, meaning little air volume. I want to saturate that small volume and then exhaust it using dry replacement air of obviously equal volume. The flow rate has to stay small.
I may end up with an air source of say 5X the heating chamber volume of already dried air just sitting there to be exploited. Testing will determine what the end result will look like. A small air pump circulating that air over disiccant should produce excellent dry air.
That demarcation line in the photo may have something to do with marketing/advertising aka professional lying.
2 Likes
Seems as though your approach might have the virtue of being energy efficient.
At the other end of the spectrum, fluidized bed would maybe be the most energy intensive, but from what Iāve read probably also the fastest. Maybe the best solution would be the option of doing either, depending on whether you have time to kill or are in a rush.
Iām not picky. Iād take either one, as either one would undoubtedly crush the consumer filament dryers that are on offer, none of which look like they would be very good, except on Mars. 
I looked at the various filament dryers and wasnāt impressed. I decided to DIY one and so far have the breadboard of electronics worked out and the bulk of the software written for the ESP32. I have it running a web server that I hit from my PC and it graphs out progress of both chamber temperature and humidity.
I wanted the dryer operational before I ordered an X1C so that I could have operational filament around the day it arrived. The thing that kept bothering me about the X1C was the small bed size. I ordered a Ratrig Vcore 4 500^3 so I can print the entire heating chamber in one go. I want a printer for engineering projects and the X1Cās small bed meant printing things in segments and then combining them and eventually I gave up on that idea.
A while ago, I built a gadget to make āblack garlicā and Iāll use that to dehydrate the initial spools of filament to then print a real working version of what I envision in a filament dryer. I might get the printer in August.
Be careful with the CO2 itās acidic when combined with moisture, we also used to find it would diffuse into neoprene rubber and cause it to swell, not sure about plastics but TPU might be a problem. Argon is inert so should be OK but itās very expensive (until we were making argon our cryogenic plant was running at a loss), nitrogen works just as well and is much cheaper. But of course all high concentrations of gases have safety concerns.
I recently noticed some compressors on amazon for pcp air guns that can produce as high as 5800psi! 5800/15= 387 times normal atmospheric pressure Wow, amazing. Even hand pumps can allegedly get up to 4500psi. I had no idea that people were operating air guns at those kinds of high pressures. Some of the air compressors come with air dryers and/or water separators.
Hereās an example of such a compressor that can allegedly reach 4500psi.
and here is an example of a hand pump with a moisture separator:
https://a.co/d/0co8CEoK
Some of the more basic 4500psi hand pumps, without moisture separators, on amazon costs as little as $40
At that level of pressure, most water vapor will be reduced to liquid water, which the moisture extraction then removes. Therefore, if Iām not mistaken, the resulting filtered air, by volume, should be quite dry indeed.
Obviously air pressures that high can be quite dangerous, so be aware and take appropriate safety precautions. Also, some of the equipment from China may not have adequate safety margin, so consider that as well.
https://a.co/d/0co8CEoK
I have large cylinders of welding gasses (CO2, Argon) that I could use to inject dry gas into a chamber to replace moist air. I donāt think that extreme is necessary, but I donāt actually know that yet.
The heating is to get molecules to expand and vibrate more violently than at a lower temperature. The expansion alone gets the air trapped inside the coils to become eligible to be replaced by dryer air from a desiccant chamber.
Experimentation may show that short intervals of minimal heating then cooling, minimal heating then cooling, etc might be more efficacious than using higher heat in one cycle; donāt know that either yet. It may also depend on the filament chemistry for which method is better per type.
My guess is that it will at least speed up the drying process. From what Iām reading, there is some danger of dry heating filaments for too long, resulting in them becoming brittle. So, from that perspective, the faster the better. Yet, I would agree better/faster may not justify the extra cost and complexity. On the other hand, some of the fancy pants filaments that need the most drying protection arenāt cheap, and so payback might be easy. And, once you have a good drying setup, you have it forever. 3D printers may come and go, but youāll probably only need one good dryer.
I donāt think weāll ever get consensus about the necessity of this. There will always be people who say (without proof) that it isnt needed, and I donāt have a proof that theyāre wrong.
The chemistry of all filaments relies on interlinked molecules. Heat is what can ruin those bonds and is why Iām using tiny 22W heaters to very slowly increase chamber temperature while graphing the temperature and humidity levels in the chamber air. Iāve initially allocated one hour to reach whatever temperature I set as the goal for any drying session. Iām not interested in fast drying, Iām interested in safe drying.
Each of the two heaters is separately controlled and with software, which is watching the temperature and humidity, I can turn the heaters on and off as often as I like to test scenarios.
The whole reason I didnāt just buy some $100 filament dryer is because my engineering background told me that their heating elements were way too powerful and use the idiocy of time to determine when to stop drying. They want to claim fast drying, but at some unknown cost in terms of chemistry.
Iāve spent several hundred dollars on all the tools, wire, micro controllers, pc boards, switches, relays, etc on the hunch that the current approach to drying is dead wrong and that drying has to be an iterative process where humidity level inside the chamber determines the correct approach, not a timer.
1 Like
Iām 100% interested! How will you test your theory? i.e. how will you know whether youāre right or wrong?
The one true thing that never seems to change is that at the consumer level, people buy conformance, whereas in the commercial world people buy performance. Generally speaking, consumers lack the knowledge or insight to make informed purchase decisions, especially in regards to technical products. Because they canāt tell the good from the bad, it all tends to be meh, or worse. So, in this context, a lot of people on the forum will say āI have a filament dryerā and thatās where it ends. China serves up gadgets which have that name on it, and people buy it, thinking that it must be good or else it wouldnāt be sold on amazon, or best buy, or walmart, or wherever. In the commercial world, in contrast, they care about how well their filament dryer works. They know what to measure.
Theyāre the smart money.
I recently was informed that my RatRig V-core 4 500^3 arrives some time near end of August. I wonāt pick the dryer project up again till that box is built, tested and Iāve gained enough knowledge about 3D printing, the software, etc to print the chambers I need for testing my dryer ideas.
I expect to go through multiple iterations of designs once I get actual real world data. Iāll use another dryer I built, to make black garlic, to temporarily dry any filament I need to some acceptable level. The goal is not zero moisture filament, but filament suitable for use in some project where the moisture level is as yet an unknown. My use of a 3D printer is for my personal engineering projects / gadgets I want for welding, solar energy, etc.
Youāre right about what the average person does. They are consumers that havenāt a clue about engineering, what to look for and how to evaluate products on the market from the efficacy perspective. Marketing, or what I term professional lying, is what drives their purchasing decisions.
BTW - I really like your conformance to performance analysis. Youāre spot on.
I hope itās easier to put together than a Voron is. The good news though is that having put it together, youāll always be able to fix it.
For a bit of levity, if youāve never seen it before:
I realize youāre doing a rat-rig, not a voron, but Iām guessing the experience might have parallels. The most Iāve done is my Prusa MK3, which took about 8 hours, which was OK but time that Iāll never get back.
1 Like
Building things is something Iāve done most of my life. Thatās what engineers do - they scratch their own itch.
I originally just wanted a 3D printer appliance and was looking at the X1C. As a professional software developer (Linux), my open source side got me to seriously look at what that meant. I would be ignorant of how the guts worked, on a Caribbean island where parts availability meant a 3 week wait for the ship to arrive and Iād have to rely on āsupportā that, in the real world, is usually hit or miss. My situation relies on self sufficiency so that meant a kit of known good parts where I learn as I go to produce a tool I can service myself. I looked at the Voron but the kit makers appeared to skimp on parts and self sourcing, given my astronomical shipping, import duty and other costs was out of the question.
Iāve never laid eyes on a 3D printer, spent over $2000 for the kit and expect to spend another $500-$1000 to get the extras for a complete set up. Doesnāt scare me one bit.
1 Like
For the folks looking at buying/building dryers and such, my journey with my own experiments keeps dragging me further down the rabbit hole.
What Iāve found so far is ambient humidity plays a huge role in how dry you can get filament in a dryer. Iāve now seen my base that exchanges a small amount of air for a Sunlu S2+ both add water to filament on humid days and remove it on dry days. I didnāt have my scale yet to weigh the rolls but it fits. I think low cost dryers are a waste of time on humid days and can actually humidify filament over whatever moisture it comes out of a bag with.
I am convinced that to make a dryer that can be effective any day without regard to ambient humidity that you have to do controlled air exchange with the outside air to help sweep liberated moisture out and the makeup air needs to be ādryā. The things that prop dryer doors/lids open are only effective on dry days. On humid days, that becomes a way to add water to your filament.
My scale just arrived yesterday but now I can verify weight loss and gain under different conditions but at this point Iād say if you open a roll on a humid day, you generally shouldnāt try to dry it. Instead, put it in a container/AMS with fresh/active desiccant.
Itās an equilibrium process. Filament and desiccant both want water. As they get dryer, they want water more. Just stating the obvious but as they absorb water, they donāt want additional water quite as much and what they have is easier to remove. Desiccant that has lots of water in it gets less effective at retaining more. Dry boxes need to have their desiccant regenerated whenever the humidity in them gets too high or they can become humidifiers.
What Iām trying now is using a fan to push air through desiccant before it enters the dryer to see how effective that is in practice. Depends on how quick and effective blowing air through a column of desiccant is.
1 Like
Part of what Iām considering is a chamber of already dried air to act as the source for injecting that air into the heating chamber. Iām planning on using an aquarium pump to move small amounts of air/second in a continuous loop over desiccant to get to the lowest humidity possible in a sealed chamber. A micro controller and sensors can determine when to turn that pump on and off as necessary determined by appropriate software. The pump is completely inside the drying chamber; its input and output are the same air. It provides forced movement over desiccant in a controlled way.
Make up air would be ambient air but it would be going into a relatively large volume of already ādriedā air so the injection of some amount of humidity shouldnāt be significant. The continuous cycling of the same air over desiccant should rapidly again reach some equilibrium point that is now the new low level of dryness that the micro controller, sensors and software can take notice of.
The other aquarium pump that moves ādriedā air (source) into the heating chamber (target) creates a vacuum effect in the drying chamber that will suck ambient air into it to be cycled through a maze that forces air in the maze to have large contact with desiccant. The aquarium pumps come with small one way valves that keep ambient air out and allow it in only when it detects a slight pressure differential between the input and output sides of the valve.
2 Likes
Sounds like a winner to me. Not quite a fluidized bed, but apparently the problem with fluidized beds made out of desiccant is that desiccants tend to be kinda brittle and end up breaking apart from the collisions and eventually turning into dust, and then you need an extra filter to remove the dust. Cycling through the same desiccant column over and over, without the friction, is, Iām guessing, the next best thing, since dwell time seems to be so important.
For inspiration, I found these industrial air dryers for about $1K:
The first one is clearly overkill in terms of CFM, but for $1K the second one might be about right if it were connected to a blast oven. That too would be overkill if @IslandBill 's hypothesis that slow and incremental would be even better. I havenāt scoured the universe, but on my first search I didnāt find any new ones for less than $1K. I did find a used one of unknown condition on ebay though for $350 (or best offer).
1 Like
Letās not forget about operating costs.
This is what I bought because it was the smallest pump I found: https://www.amazon.com/gp/product/B09JKHPT2S/
Run it the 24/7/365 and it would cost less than $5/year at $0.35/kWh .
3 Likes