My suggestion, and it’s only that so take it as nothing more, would be to simplify the directions as much as you can. Maybe link to more detailed information that you’ve written for those who want it, but leave that as optional. Otherwise I’m afraid the impression people may get (especially when compared to other makersworld projects) is information overload. In general, people like simple, so the more simple you make it look, the easier it is for people to get on board. If you wanted to do a youtube video on the build, I’m sure that would get a lot of people over the hump, plus draw in the much larger youtube audience too. But these are just suggestions on what I might do if I were in your shoes. I think you did a great job building it, and now it’s just fine-tuning the marketing.
That may just be the nature of things. I was once a prolific uploader of PCB designs to a particular home automation forum, and I had literally thousands of downloads, which were all free. The fact is very few people ever follow up with a question or even a “thank you.” You’ll notice the same kind of thing on github. So, it doesn’t mean none of the 13 built your dryer. It may simply be that out of those who did, they’re simply lacking the time/energy/will to post about it. That’s probably more typical than not. So if you do it, do it because it’s something that you want to do. For the most part, with some exceptions, the world will act like it doesn’t care one way or the other. You may be having an impact, but odds are you’ll never have confirmation as to the extent of it.
Probably the best payoff is that it’s a beacon that will attract like minded people to you, who share your interests, and those are good friends to have that you might not have met any other way. 
So, for that reason, I still think it’s a worthwhile activity. You do the work anyway, for your own reasons, and writing it up afterward and posting it doesn’t take all that long. It’s a bit like writing a message, putting it in a bottle, and throwing it in the ocean. It may or may not land somewhere or get read, but sometimes it does, and it’s fun when that happens. Sometimes years later even.
I think you are correct on the complexity stuff. It is intimidating but two things led to that - the rise of people downrating models that are beyond their capabilities, and a lot is leftover from the earlier versions where there was all the gluing with lots of ways to go wrong.
I think I will trim it down, though. Thanks for the critique. The only trouble spots really now are painting it to seal it and soldering to those hygrometer battery terminals. The funny thing is the last one I printed was the tightest to air and best finish, etc, I was using filament that was nicely dried. So those without dryers but having moisture issues will likely see the same thing - if they print an air dryer it could be leaky. Kind of ironic.
Reporting back: Unexpected, yet welcome result. It appears that running the equipment this way, with light positive pressure, may have made the air within the drybox gradually even dryer. I’d want to change the TH sensor’s position within the drybox to confirm, but that is the preliminary result.
Good lord when you two get going it get’s harder to keep up.
When you’re done can I get Cliff Notes?
So the positive pressure slows down the effective draw out of the water. I guess that makes sense, since I would expect a negative pressure to speed up the effective drawing out of water.
Update: Today I purchased a vacuum chamber and vacuum pump. Prices have dropped a lot since the last time I owned such gear. The vacuum pump that comes with it claims that it can pull a vacuum of a mere 37.5 microns., which, if true, is far better than the vacuum pump I previously owned and far more capable than what CNC Kitchen was able to attain with his vacuum pump.
Although it is an oil based vacuum pump, which I would ordinarily dislike because of all the smog it produces in the work environment, I just today became aware that there are “anti-misting” oils that are available for use in vacuum pumps… This should eliminate the smog problem; at least I’m hoping so. I’m going to give it a try and see how it goes. I’m thinking that the vacuum chamber should also be handy for drying silica gel desiccant desiccant without damage to the indicator dyes. If I’m lucky, it may also allow me to dry molecular sieves without needing a muffle oven to generate the normally recommended very high molecular sieve regeneration temperatures. For the moment, this is just a theory, but if the vacuum pump can really pull 37.5 microns, then… I’m hopeful!! On its face, it seems worthy of some FAFO. 
I also ordered a digital manometer (surprisingly cheap!) for more precisely setting the minimal positive pressure in the above “active drybox”. At the moment I’m leveraging an oxygen flow sensor (as seen in the photo) to help me accomplish that, but it can’t account for leaks, whereas a manometer can. The same manometer will also help me measure the amount of vacuum in the vacuum chamber more precisely than the analog dial gauge that’s part of the packaged vacuum chamber kit.
As these various pieces fall into place, I’ll hopefully start making monotonically faster progress in 1. drying filaments more quickly, 2. drying them more thoroughly and 3. keeping them dry in an active drybox throughout the time they’re used or needed during the 3D printing process, no matter how long that might be. Then I’ll be better prepared to get really good prints involving nylon, TPU, PPS, ABS, ASA, and any other moisture sensitive filaments that may come along or is still yet to be invented 
I hadn’t even thought of that, but I’m doubtful it’s the reason. Why? “positive pressure” may sound like it must be powerful, but in this case it’s incredibly weak. I’m deliberately setting it as low as I can, because it’s only purpose is to make sure that air leaks out of the drybox as a way of guaranteeing that no ambient air ever leaks into the drybox and thereby prevents any ambient moisture from intruding into the drybox (for as long as positive pressure is maintained).
“But if not that, then what?” I hear you asking. My hypothesis is: since I haven’t yet finalized arrangements to use dry make-up air while drying my printer’s filaments, the filament in the drybox–even though I did dry it conventionally using heat and ambient make-up air–never got as dry as it theoretically could have under more ideal circumstances.
In other words, the heated make-up air used to dry the filament had a particular dewpoint, and the filament dried until it’s remaining moisture content became in equilibrium with that dewpoint. At that point, it could dry no further under those conditions, and therefore the drying cycle ended. Let X be that dewpoint.
Now, currently, with the filament in the drybox, I’m pumping in desiccated air to maintain positive pressure, as already described. Let Y be the dewpoint of that desiccated air. If Y is a lower dewpoint than X, then my hypothesis is that as the Y “positive pressure” air washes over the filament on its way to being flushout out of the drybox, it may carry with it some of the remaining moisture still contained in the conventionally dried filament. It’s a very slow flow of desiccated Y air, so the effect happens only gradually, and slowly at that, which is what we see in the timeplots I posted just above.
You make quite a few assumptions, but so do I. Have you considered the effect the pressure difference has on your measuring device (hygrometer)? We know that changing the pressure moves the dew and boiling point of water, but that is not the only effect it will have.
My concern would also be what else it is effecting. You measure quite a change with your modest positive pressure. Was the change only in the water? Could it have affected the hygrometers ability to measure accurately if it is designed to measure at normal pressure?
I really like where this is going, but lets remember it is not just water that will be effected by the pressure.
Well, yes. All valid points I suppose. It might even be the case that it’s as simple as predominantly the progressive dilution of one “solution” by another, where in this instance one solution is the native air and moisture mixture inside the drybox and the diluting solution is the dryer air that’s being pumped in.
I’ve since thought of an experiment which can check to see whether I’m merely measuring my own dilution, and maybe when time allows I’ll do it. There’s just so much to check and cross-check and confirm that it’s all a bit much for any one person to accomplish in his free time, especially for someone like me who is largely feeling his way through the dark. That’s FAFO for you. I’m sure someone better versed in the basic physics/chemistry of the subject matter could make short work of the topic by leveraging textbook knowledge. Come to think of it, I’d probably get better traction posting this on a physics forum instead, as this entire endeavor would likely be child’s play for them. Now that sparky has moved on to other projects, there’s no real quid pro quo for continuing it here.
So, with that in mind, I’ll leave the topic with this fun video:
which demonstrates why such a hard vacuum is beneficial, and why I should return my weaker oil-free laboratory vacuum pump before I even unbox it. i.e. evidently at ambient temperature, you can dry silica gel desiccant at below 1200 microns, which is still quite a strong vacuum, but get little to nothing at a weaker vacuum. This may in part explain why CNC Kitchen had such lousy results in his vacuum drying attempt. His gauge showed that he wasn’t drawing a vacuum anywhere near as strong as that, as seen in this screenshot that I clipped from his video at the point of his strongest vacuum pull:
Granted, whatever filament he was testing isn’t identical to silica gel, but for all we know maybe it’s loosely comparable in its attraction of moisture.
The guy in the video was only able to draw 200microns of vacuum, despite having a pretty monster looking vacuum pump. The vacuum pump that’s coming as a package with my vacuum chamber (<$90 for everthing, chamber and pump both, including shipping) is claiming it can do 37.5microns. I’ll check when the digital manometer arrives. For the money, I’ll be impressed if it can really deliver it. I hope it isn’t mere puffery!
One last thing worth mentioning, which I only just recently learned about, is that virtually everyone with a vacuum pump who is using it to remove moisture is, essentially, abusing their vacuum pump. That’s according to this guy, who seems to know quite a bit about the subject:
Luckily, I’ve never before been so well equipped to deal with precisely this issue, and so I intend to dry the air as it gets pulled into the vacuum pump. On the other hand, if I’m using desiccantg to dry desiccant, then it may be inherently counterproductive,because it may well put me directly back at square one with respect to drying desiccant . Nobody ever seems to ever talk about that aspect of it: on its face, I would just be shuttling moisture back and forth between two banks of desiccant. Right?
All in all, utilizing only a vacuum pump and no added heat as the means to dry desiccant strikes me as unpractical, at least if the earlier video demonstration is typical of how long it takes. I may change my min on that, though, if I’m able to pull the promised 37.5 microns and if that results in a much, much faster dry time.
- The vacuum pump I’m receiving has a duty cycle of 30 minutes before needing to cool down.
- A lot of vacuum pumps recommend an oil change after a cumulative 12 hours of use, though I’m unsure whether that may be related to the likely water buildup in the oil (see youtube video directly above for further information relating to that). Maybe if I dry the intake air that can be extended by a lot, but I need to learn more before I can say for sure.
Rather than do all that, it may, in the end, turn out to be more cost effective to buy silica desiccant in bulk and then dump it rather than recharge it as it gets used up. You can buy 50 pound sacks of the stuff for cheap, and I can only imagine that 50# of it t would surely last quite a long time.
Nah, it would be cheaper to Nuke the the silica dessicant, instead of buying more 
Reporting back: I couldn’t find recommendations on miminum dry air pressure into a drybox such that the positive pressure would guarantee no moisture from the ambient environment would infiltrate in, but I did find some comparable information regarding cleanrooms, and you may be surprised at just how little positive pressure is needed. In the case of cleanrooms, the range is 5 to 20 pascals How much is that? Well, converting to psi, 5 pascals is 0.000725 psi. Barely anything! So, if you were to dial-in the amount of dry air entering the drybox all the way down to that, I’m guessing that the desiccant column would last a long, long time. Maybe with some assumptions and further calculation we could even estimate how long.
My more immediate concern was how to measure such a low pressure so that I can dial it in, as this is undoubtedly far below what a typical flow gauge is going to show. I ended up with the Testo 510i, which can allegedly measure pressure differences down to 1 pascal. I expect it will arrive from China in about another week. It costs about $100 on aliexpress.
In retrospect, if I had calculated the dry air flow rate needed to sustain 20 pascals of pressure, maybe I could have simply counted bubbles to set the flow rate while dialing it in. Too late for me, but for anyone else, that would be a cheap way to do it.
Even though this approach should keep moisture laden ambient air from wafting its way in, I suspect it wouldn’t relieve the osmotic pressure for moisture to find its way into the enclosure anyway. So, maybe a higher flow rate will be needed to flush that out fast enough that it has no impact. That’s a calculation for another day. Just not today. But for that reason maybe the best choice for a drybox would be either all metal or all glass–something truly impermeable, where the only point of vulnerability would be a gasket or o-ring. Or, if plastic, then a thick enough plastic that the infiltration rate would be very, very low.
It looks like most of the comments are on the cost of the freeze dryer or the difficulty of making your own. I recently purchased a Harvest Right for food preservation and not for drying filament and the same idea occurred to me. This is still the most detail post I’ve found on the subject. My concerns on the process are:
- a spool will not fit on the standard trays
- If tray is removed will the device function without the temperature and sensor umbilical from the tray.
The way this would have to be done is that the filament is removed from the spool to some sort of rig to keep it detangled and then respooled when complete.
Let me know if anyone hast tried this yet.
I do wonder if there are any side effects to this, relating to the material strength of the filament.
I shot a message to the company that makes my freeze dryer (Blue Alpine) and they responded quite quickly stating that they use their freeze dryer to dry their filament. Some of their parts are 3D printed so makes sense. They said it works quite well. I requested they do a YouTube video on it so we’ll see.
If you go this route, be sure to get a vacuum pump whose oil is easy to change, because from what I’ve heard some amount of the extracted moisture ends up there, necessitating a relatively frequent oil change (at least with the typical Harvest Right design), though I can imagine ways that might be mitigated. There’s one vacuum pump in particular that makes oil changes nearly a push button process.
Thanks, @NeverDie. That pump looks pretty slick. The pump that Blue Alpine includes is really easy. Not 1 button easy but you just loosen an Allen key plug and drain.
@Not_Neil_Armstrong Please do let us know how it all works out for drying filament if you decide to try it.
This guy is long-winded, but he gives a pretty good explanation of the details of how a freeze dryer actually works:
For instance, prior to this video, I didn’t realize that much of the extracted moisture is trapped by freezing to the metal walls of the chamber, or that the trays are heated at a later stage of the freeze drying process by heated silicon matts that heat the metal trays you insert into the freeze dryer. The latter makes sense, because how else could you heat them in a vacuum? However, they aren’t visible in most photos, so you likely wouldn’t know unless you owned one.
@Not_Neil_Armstrong Also worth trying if you get one: rewind your filament onto a metal spool prior to the freeze drying process. Maybe the improved heat conduction from those silicon pads could radically reduce the drying time.
Speaking of which: you should ask the blue alpine guys how long a typical drying cycle takes in their freeze dryer. When it comes to food, the guy in the video was saying it was something like 16-24 hours, with the vacuum pump running the entire time. Since there’s obviously a lot less moisture in a filament spool to begin with, maybe it’s a lot less. But who knows for sure. Well, it sounds like the Blue Alpine people could say for sure if they’re doing it already.