How are people keeping their filament dry?

Bambu Lab A1 Series Bambu Lab A1 mini
143

#1: Great idea. Agreed. I presently own three different models of of filament dryer, and one of them is the Print Dry Version 3 filament dryer, which, for what it is, is a stupidly expensive piece of kit (around $199). At the time I purchased it, I was in a rush and so I relied on what seemed like near consensus among the reviewers that it worked decently well. Moving along: I put two different wireless TH sensors inside PrintDry while drying some PLA to monitor the conditions inside. Both were made by Yolink, but they were different models and both were specā€™d to have an operating temperature range up to 70C (by their own measurements they were comfortably under that limit). Naturally, they did report a lower RH than the ambient RH in the room housing the Print Dry, as you would expect because of the much higher temperature inside the Print Dry while drying the PLA. As a convenience, both sensors also report Dew Point (which if they hadnā€™t reported, you could easily calculate using either formulas or an on-line calculator, given the measured temperature and RH).

Worthy of note: there are a lot of holes in the PrintDry ā€œshellā€ which serve as ā€œportsā€ for filament inside the Print Dry to connect to a 3D Printer located somewhere nearby. For this reason, I was initially worried that the dew point inside the PrintDry might rise enough to, in the limit case, go as high as the Dew Point in room housing the Print Dry filament dryer. Instead of that, however, both sensors reported that the dewpoint inside the PrintDry was actually higher than even the ambient dewpoint in the room housing the Print Dry. And this was after even many many hours of continuous drying inside the Print Dry. WTF? If the PrintDry had been airtight, this would, of course, make sense. But for a filament dryer that leaks like a sieve, as seems true for the Print Dry, I was surprised by this result. I have a couple different theories as to why this might have happened, but regardless of why it happened, it would maybe support your idea of pumping already dry air into the Print Dry during the drying process so that it could dry faster and more efficiently. Indeed, IIRC, BBL recommends doing that too, except upping the ante even further and doing it in combination with using a blast oven (well, at least for certain kinds of filament).

So, #2, not only is it a good idea, but for at least some situations, it may be highly advisable, and maybe even a necessity for some kinds of filaments.

Finally, with that background now covered, we can cut to the chase: #3, how best to do it? Well, lots of ways, but how to do it easily and economically? Thatā€™s the real question. Iā€™m sure using enough tanks of nitrogen, or scuba tanks as some suggested, would get the job done, but surely there must be an easier and more cost-effective way than that?

144

Use two ports, one in and one out, to inject dry air and expel moist air. The injection will lower the temperature, but the heater should be able to compensate rapidly, especially if you use a low volume air pump. Since drying is expected to last for quite some time, thereā€™s no need to pump massive amounts of air per unit time.

Iā€™m just reading the chamber humidity level and when it wonā€™t go any lower over a certain time period, the drying is done, dispensing with dew point calcs.

I purposely wait till the chamber humidity stops rising before I pump in dry air and stop when it stop falling to reach a new equilibrium point. Wait for a new rise to a new max, pump in dry air, wait for a new low, ā€¦ repeat till done.

145

@IslandBill I was a little less than half asleep when I wrote my post earlier today, but unfortunately I was a little more than half asleep when I decided to go back and edit it. Unfortunately, I started my edit before you replied, but I didnā€™t finish my editing session until after you replied, and it wasnā€™t until then that I saw you had replied at all. So, apologies, Iā€™m afraid I may have muddled the discourse a bit. Iā€™m back now with a new post to hopefully straighten it out.

Basically, yes, I understand what your driving at in your last post, and I agree with it. In fact, I think it may pretty much have to be pulsed in the way that you described, especially if attempting to dry a low temperature filament like PLA in an otherwise high humidity setting.

My question basically: whatā€™s the best way to source that dry air that you pulse into the filament dryer? I donā€™t have scuba tanks myself, and Iā€™d rather not buy bottled nitrogen for this purpose (the way Costco does when they inflate your tires), so I hope thereā€™s an equally viable but cheap way to dry out the air before pumping it into the filament dryer. For instance, IIRC, donā€™t some shop air compressors employ a means to dry the air before compressing it (so that water doesnā€™t pool within the tank, possibly rusting it out)?

Alternatively, I know there exist continuous duty-cycle commercial dehumidifiers that use desiccants on a wheel to dry the air to a much drier level than common refrigerant dehumidifiers do. The idea seems simple enough, and we could make do with many fewer cfmā€™s than those commercial dryers, so maybe something along those lines would be cheap yet equally effective?

Anyone have any concrete suggestions or pointers on particular products to look into?

Maybe even a fairly simple buffer tank of some kind, perhaps filled with molecular sieves, would do the job in a batch processing mode. In that kind of scenario, you might have to swap tanks between uses and dry out the one that has absorbed enough moisture from the air that it needs to be recharged.

146

The plan in my head is for a 3D printed tower, either round, square or rectangular with multiple ā€˜floorsā€™ where each floor has X amount of desiccant in a spiral (round, square, rectangular).

On the 1st floor, thereā€™s the air being pumped in via a fish tank pump, a tiny one, to allow time for air/desiccant contact to occur. Between the 1st and 2nd floor thereā€™s one shaft that allows air thatā€™s gone through the 1st floor spiral to reach the 2nd floor spiral that forces the air to run over all the desiccant there. From the 2nd floor to the 3rd floor, thereā€™s another shaft, wash rinse repeat for as many floors as is necessary to give you relatively dry air considering your environment.

Admittedly, I havenā€™t built this since Iā€™m waiting on my printer to arrive. It may take looping a quantity of air through the floors multiple times to get it ā€˜dryā€™ or lots of floors with each floor just high enough to support X thickness of desiccant to ensure air/desiccant forced contact. Each ā€˜floorā€™ has to be sealed off from the floor below so that the air can only rise up through the shaft and that means a seal of some kind between floors, possibly a soft plastic (TPU, etc).

The top of the tower is the exit for conditioned dry air to be injected into the heating chamber. Each ā€˜floorā€™ is removable to exchange desiccant as needed.

147

@IslandBill Notionally speaking, maybe a desiccant cannister like this:
https://www.printables.com/model/416657-ams-bambulab-adapter-give-weight-and-include-desic
which is closed on the sides but open at each end. Then stack a number of them end-to-end together in a tube or pipe of some kind to hold them tight against each other. Ambient air flows in one end, and dry end out the other. If thereā€™s too much resistance to airflow, you increase the diameter. It would be much like your ā€œfloorsā€, but sans tubing between them. When itā€™s time to recharge, you could pull them out, bake them in an oven to regenerate the desiccant, then re-insert and be good as new.

Iā€™m not sure whether slice engineeringā€™s desiccant cannister is open on only one end or on both ends, but if on both ends maybe it could serve the purpose:
https://a.co/d/47zbquR

I like that their ā€œactivated aluminaā€ regnerates at 250C, because then you can bake it really, really dry, even if your ambient humidity is high. On the other hand, silica gel has a maximum recommended regeneration of 130C (before it starts releasing toxic chemicals), so maybe that temperature (or a bit less for a safety margin) would still be good enough.

Which would be the best desiccant to use? Slice claims its activated alumina is 10x more effective, but I do wonder if thatā€™s on a gram for gram comparison basis, or volume to equal volume basis. Maybe if you threw 10x the amount of silica gel into the problem, youā€™d get better airflow in a wider package (since less resistance) and overall humidity reduction would be the same? Anybody know? This drying tower would probably be located outside the filament dryer anyway, so seems like thereā€™d be plenty of space to do whatever you wanted.

148

I had to build a custom (forced) dryer for a project at work a few years backā€¦
Not sure what route your are going but maybe this rather simple design concept can give you some additional ideas:

We needed a half decent, yet cheap way, to ensure a manufacturing chamber for a machine would not get above 30% in humidity.
Two main problems:
a) The machine would go to a rather less developed part of Oceania where, for most the year, humidity levels and temps are high high.
b) It had to be low tech in order to be able to repair and service it without the need of imported parts.

The solution was a set of 100mm PVC piping, a simple refrigeration unit, a microwave and a few kg of silica gel beads of the indicating kind.
A fan forces the air out of the chamber and into the PVC pipe where it goes through the Condenser of the tiny aircon.
This heats the air up and then it goes through the Evaporator of the aricon.
The hot side was kept moderate in terms of heat generating (oversized) while the cold side was optimised for a surface temp just above freezing.
An expansion valve with heat bubble mounted to the cooling side was used to ensure hassle free operation.
The condensed water was allowed to drip off and collected to go out through a tiny hole and drain hose.
The final stage before entering back into the chamber was a section of pipe filled with about 1kg of silica beads.
Every morning the beds are replaced and ā€˜rechargedā€™ in the microwave.
We supplied like a spatula to glue onto the ceiling of the microwave in order to stir up the beads while the plate is spinning.
Donā€™t remember the settings but think it was like 40 minutes on defrost to dry the beads - or until the turned back yellow again.

The goal was to get around 30% of humidity even on a rainy day.
Once operational it managed to get to around 25%.
For 3D printing I would probably have the Condenser outside the system as no additional heating for the chamber is required.
If it is for engineering plastics though it might be beneficial to use a split Condenser so the air wonā€™t be cooled down (as the average).

As for the silica get beads:
Go for the largest size out there.
We used 6mm beads.
Those tiny ones we find in the satchels and bags cause too much restriction if just a normal impeller fan is used.
To further reduce the restrictions it really helps to have large diameter pans sitting in ducts.
Remove the lower, discoloured ones for recycling, drop the rest down and put fresh ones on top.
Could be done for under 100 bucks in parts from the hardware store and a salvaged water cooler from the hard rubbishā€¦

2 Likes
149

Thatā€™s a fairly elaborate set up and extra energy intensive. One of the few things Iā€™m not capable of repairing myself is anything to do with refrigeration; Iā€™ve just not studied it and have no tools related to it.

I live in the Caribbean where itā€™s always above 70F and usually above 85rH. The printing will be done in non A/C space inside a warehouse where I have an A/Cā€™d office. I donā€™t want the fumes and particulates anywhere near my lungs, so itā€™s going to live in a somewhat hostile environment while operating. When not functioning, it will be in A/C space living on a purpose built rolling cart that will also house canisters of already dried filament.

Iā€™m going to try my less involved plan to see how it goes. I figured that at worst, circulating a given and static volume of air inside a chamber full of desiccant should reduce itā€™s humidity to an acceptable level. The question being how quickly will that happen. That storage volume of ā€˜dryā€™ air could be the source for periodically pumping some into the filament drying chamber. Make up air would pollute the already dried air to some extent but if the circulation is constant and the need for dry air isnā€™t, there should reach a balance that I suspect should be acceptable.

Time will tell. This is an experiment.

150

Those two examples are tiny compared to what I have in mind. Iā€™m thinking of an air dryer about a cu ft in volume as I described having multiple ā€˜floorsā€™ to expose lots of desiccant and spiral channels per floor to force the air to be turbulent over a restricted portion of the floor as it snakes around the spiral to reach the shaft up to the next floor.

Itā€™s the air to desiccant contact that I think needs to be maximized for rapidly drying the air and the volume of dried air being a somewhat stable mass of available air to feed into the heated filament drying chamber every few minutes. I have a huge air compressor that has an air dryer on its outlet and its full of beads with a huge volume of air rushing through them and it works fine for spray painting and other tasks where water would be a problem.

Iā€™ve searched the 3D printables type sites and found nothing even remotely like what I intend to draw up and then print.

I should mention that I decided to not purchase a Bambu product and am awaiting delivery of a Ratrig V4 500^3 that will support the printing of both the drying and heating chambers in one go each. The Bambu support issues Iā€™ve read about plus the fact that the word ā€˜warrantyā€™ means nothing where I live since shipping anything into and out of the island costs a relative fortune, were deciding factors. I reasoned that I needed to build the printer to be able to self maintain it afterwards instead of buying an appliance I know nothing about and need support and parts from a single source. Being a Linux developer and supporter of open source also aided in my decision.

151

Iā€™m using a Sunluu S4 to dry the filament then bagging them un Sunluu vacuum bags with a tub of silica desiccant.

Expanded my collection of filament so my use up rate is a lot slower, so went got a bit more serious.

152

I have food storage containers waiting to individually house a dried spool along with desiccant and a cheap humidity gauge.

I may decide to install humidity sensors in the dozen containers I purchased and tie them all to an esp32 that can send me alerts when any particular container starts approaching a problem state. Experience will determine future approaches.

153

This may turn out to be very easy to do if there is a desiccant which can drive RH all the way down to 0% in a confined space at an air temperature of 60C. I know from running some simple experiments that at 20C (normal ambient temperature inside most human dwellings), I can drive RH down to 0% in a sealed container with regular fresh silica gel beads if I use enough of them. Is the same true at 60C? If so, then problem solved: Put the PLA spool in a sealed box with sufficient fresh desiccant in it and with a heater-fan in it that will raise the air temperature inside the box to 60C (to drive the moisture out of the PLA spool). The desiccant will absorb the moisture and drive the RH down to 0%. After spending enough time in the box, the spool of PLA will be totally dried out.

Will that work with silica gel as the desiccant? I donā€™t know, but if not, then Iā€™m optimistic that one of the other available desiccants would do the business to get RH down to 0% at air temperature of 60C.

154

In a sealed environment, trying to move, not eliminate, moisture is only going to go so far. I think itā€™s necessary to expel as much moisture as possible using heat to excite and expand the molecules out from hiding in the air trapped deep inside a spool. The hot environment is not conducive to how the desiccant would like to work and therefore can only go so far during a heated drying session. Thatā€™s why the software Iā€™ve written to date senses for a low in humidity and terminates the drying and doesnā€™t use time at all. This should minimize how long filament is treated poorly just to get rid of moisture.

Thereafter, placing the somewhat dried spool into a room temp environment with desiccant over an extended period of time in a sealed container should allow Brownian motion and changes in atmospheric pressure to acquaint every air molecule with the desiccant now at a more appropriate temperature for that desiccant to further function.

It all depends on what your ambient environment is like. Las Vegas would be vastly different than Key West.

1 Like
155

I think weā€™re actually all in violent agreement with each other. The main question is over what air temperature range is any given desiccant still effective at drying the air RH down to 0%? All of us know for sure that it can work at 20C, and if it turns out that no desiccant is effective above that temperature at driving RH down to 0%, then we need these elaborate schemes. All Iā€™m saying is that if it turns out that one or more of the possible desiccants, in sufficient quantity, will drive RH down to 0% when the air temp is at 60C, then you can avoid that complexity by selecting the right desiccant in sufficient quantity and using the sealed box approach. Which would be easiest to implement?
If either method would work, thenā€“at least to meā€“sealed box would seem to be a whole lot easier. v Worthy of note: With a sealed box, you might need a simple one-way pressure relief valve because of Boyles law and probably not wanting to create a pressurized environment.

Itā€™s a simple enough question. Maybe you guys already know the answer. Iā€™ll chase it down myself, and then Iā€™ll know for sure. Heck, maybe even chatgpt could answer this question.

156

I know desiccant works in my huge air compressor thatā€™s locked in a 20ā€™ shipping container on an island at 16 latitude in the summer where the interior temp is somewhere around 45C and humidity is 85+. The roof of that container softened the soles of my sneakers.

157

Reporting back: According to Desiccant Types - SorbentSystems.com
silica gel is less effective at the higher temps, but molecular sieves should work quite well anyway well past 60C.

On the other hand, the same source says silica gels will only reduce RH down to 40C even at normal ambient temperatures. WTF? If I put 500g of fresh silica gel in a sealed cereal container with a wireless TH sensor inside, the TH sensor shows RH drops to 0%. Iā€™ve done this many times and with different wireless TH sensors, so I donā€™t understand why it would say it only drops to 40% RH.

Edit: Anyhow, for a quick proof of concept, one could use a pelican case, as those are air tight and have an air valve built into them already (for adjusting to different altitudes). Even amazon sells molecular sieves, so acquiring those will be no problem.

I enjoyed the brainstorming. Good luck guys with whatever it is you decide to do. I just want to nail this once and for all and be done with it so I can move past this issue and get on with my printing.

158

By the way, for anyone who might be interested in the pulsed method filament dryer, hereā€™s a DIY build that illustrates the basic idea:

In the video you can even see the moisture building up inside and then getting blown out with each pulse. In this particular example, the make-up air during the pulses is sourced from the ambient environment, which in some instances may work well enough if you live in a dry enough climate, but would be self-defeating (or at least self-limiting) if you lived in an extremely humid climate. In that case, youā€™d want to source dry air (as described by @IslandBill above) for the make-up air injected into the chamber during the pulses. Assuming moisture/humidity doesnā€™t leak in through other pathways, then after enough pulses the moisture in the filament will have been removed from the filament and evacuated out of the chamber, and at the end of the process the filament spool will be left sitting in dry sourced air.

There is a elegance to this approach, and depending on the source for the dry make-up air, it might even be less work and less hassle over the long-haul. So, for that reason, I remain interested in it, especially if someone else (not me) wants to take the lead on developing it. As fun a project as it sounds, I canā€™t be point man on the pulsed approach because I have near-term needs that are going to force me to take the shortest possible development path for getting adequately dry filament. I expect I can throw together the sealed/valved box approach without much effort and when heated up with the right choice of desiccant it should ā€œjust work.ā€ Or so I hope. If it fails, Iā€™ll be right back here at square 1, at which point I might move ahead with dry nitrogen for the make-up air, which should surely work, just to ā€œget it doneā€ in the near term and as a bridge to some kind of better, future long-term solution.

1 Like
159

I just realized itā€™s even easier than that, provided you are using a filament dryer that has enough extra room in it. Essentially this: run it as you normally would until your filament stops getting any drier (which you can determine by weighing your filament spool and noticing when it no longer loses weight). Then pack in the high temperature desiccant and seal it up and run it even longerā€“assuming that doing so is safe for whichever particular filament dryer you have (I canā€™t speak to every possibility thatā€™s out there). Bingo, bango, that should do it.

This way you may expunge most of the moisture from the filament without wasting desiccant on it. You only use the desiccant once the ambient humidity in the make-up air limits driving out the last of the moisture from the filament.

160

Iā€™m an absolute beginner so if someone has tried this and it has not worked, please let me know. I use 1 gallon zip loc bags. I take the spool off the printer when it is not in use and put it in the bag along with a desiccant pack, close the bag and put the spool in the original box. The spool is never left on the printer if it is idle for more than an hour or two. Neat, clean and easy. I live in the Ozarks and humidity in the house can be 60%.

161

Iā€™m sure someone has mentioned it, but I use a Sunlu S4. 2 slots next to the printer usually have ABS stored in them, The other two will be for either storing more ABS or drying whatever I need dried. Have an S2 as well but donā€™t really have room for it. The S4 can maintain humidity for storage or just power up for x amount of time for drying.

Filled filaments or TPU can be fed through PTFE tubes directly to the extruder. Depending on the amount of printing done I will switch dried ABS to the AMS. Currently printing a 11 hr print in TPU.

Long term storage goes in vacuum bags.

162

I keep a sealed box from Walmart with all my filament and several dessicant bags